WO1997000575A1

WO1997000575A1 - Object recognizing device and image pick-up device

Info

Publication number: WO1997000575A1
Application number: PCT/JP1996/001700
Authority: WO
Inventors: Yujiro Ito
Original assignee: Sony Corporation
Priority date: 1995-06-19
Filing date: 1996-06-19
Publication date: 1997-01-03

Abstract

An area detecting circuit (38) receives all picture element data about image pick-up signals from an encoder (37) and selects the area in which the picture element data coincident with the color of a preset target object exist, from among (128) divided areas. A CPU (4) receives the picture element data existing in the selected area from a frame memory (39) and prepares an object information table showing the coordinate position of the object having the picture element data of the color coincident with that of the object. Then the CPU (4) determines the present coordinate position of the object based on the data related to the preceding field stored in a target object history table. The CPU (4) supplies to a focus control circuit (34) an offset value used for changing the center position of a window for detecting an evaluated value from the data representing the coordinate position of the object. The circuit (34) calculates a plurality of evaluated values from a newly set window and generates a control signal used for controlling focus.

Description

Specification

Title of the Invention Subject Recognition Device and Imaging Device

The present invention relates to a subject recognition device for automatically recognizing the position of a subject photographed by a video camera or the like, and to automatically track a subject by using the image recognition device, and to focus on the subject. The present invention relates to an image pickup apparatus that adjusts the focus. Background art

Conventionally, in consumer video cameras, an auto focus that dynamically adjusts the focus to the subject has been used.

In general, it is well known that in order to detect whether or not the focus is correct, it is only necessary to determine whether the contrast of the imaging signal is high or low. In other words, the focus is high when the contrast is high, and the focus is off when the contrast is low. By extracting the high frequency components of the image signal, generating data that integrates the high frequency components present in the specified setting error, and using this integrated data, is the contrast high? We can judge whether it is low. This integrated data is data indicating how many high-frequency components exist in the setting error, and this data is generally called an evaluation value. Therefore, the autofocus can be realized by driving the focus lens such that the evaluation value is maximized (that is, the contrast is maximized).

The evaluation value extracted in this way is not an accurate evaluation value that represents the contrast of the subject. There are several possible reasons for this. One of the factors is to detect the evaluation value Must be set for the target subject, but when the target subject is moving, etc., there is a problem that the evaluation window cannot be set accurately for the subject. Was. For this reason, accurate evaluation values could not be obtained, and it took a very long time to adjust the focus force.

For example, in a video camera device used for a broadcasting station or a business, a photographed video may be transmitted to a home through live broadcasting. If it is not possible to obtain an accurate evaluation value as described above during such live broadcasting, the autofocus operation takes a long time. As a result, the blurred image signal is transmitted to the home. Therefore, video cameras used for broadcasting stations and business use do not require any simple, inexpensive, or compact autofocus devices, such as consumer video cameras, and have high precision. Focus control and high-speed focus control are required.

On the other hand, as a conventional technique, an imaging apparatus having a function of automatically following a specific subject photographed by a video camera, that is, a so-called automatic tracking function, has been proposed. In these conventional imaging concealment methods, first, the color components of the subject to be automatically tracked are stored, and the imaging data obtained for each field or each frame is analyzed, and the subject to be dynamically tracked is analyzed. Searches for the position of the color in the screen where the image was captured. At this time, it is necessary to analyze all the imaged pixel data in order to find out where in the imaged screen the color of the subject to be automatically tracked exists. However, if an attempt is made to perform an analysis process for each pixel of all captured pixel data, a general-purpose micro computer is used unless a special micro computer with extremely high computational performance is used. Image data could not be processed in real time using a computer. With a micro computer, In order to perform the processing in real time, the obtained source image data is decimated in the horizontal and vertical directions to reduce the amount of image data. Conventionally, the processing load on the system has been reduced. However, when decimation processing is performed on image data, the number of pixels that are processed by the micro computer is greatly reduced, so that the position of the subject to be automatically tracked can be accurately determined on the imaging screen. It was difficult to grasp.

An object of the present invention is to accurately and real-time grasp the position of a target subject and obtain an evaluation value corresponding to the position of the subject. Disclosure of the invention

A feature of the present invention is that in an object recognizing device for recognizing the position of a target object, an image pickup means for outputting an electric image pickup signal, and a plurality of image pickup images constituted by the image pickup signal from the image pickup means. Area searching means for selecting an area in which pixel data having the same color component as the target object exists from the divided areas, and the imaging means Storage means for storing pixel data corresponding to the imaging signal from the apparatus; and pixel data corresponding to the area selected by the subject-to-be-searched search means are read from the storage means, and the read pixel data is stored in the storage means. And a processing means for calculating the position of the target object based on the object recognition device.

A further feature of the present invention is that, in an imaging apparatus for imaging a target object, an imaging device that outputs an electrical imaging signal, and an imaging screen that is configured from the imaging signal obtained from the imaging device. To set the detection window, detect the evaluation value of the imaging signal in the set detection window, and control the focus based on the evaluation value. Storage means for storing all pixel data corresponding to the imaging signal from the imaging means, and calculating the position of the target subject based on the pixel data read from the storage means. An imaging device including processing means for controlling a force control means such that the position of the detection window coincides with the calculated position of the target object.

¾ BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 is a drawing showing a full rest configuration of an image pickup apparatus composed of a video camera.

FIG. 2 is a drawing showing a specific configuration of the auto-force control circuit 4. '

FIG. 3 is a drawing showing a specific configuration of the horizontal direction evaluation value generation circuit 62.

FIG. 4 is a diagram showing a specific configuration of the vertical direction evaluation value generation circuit 63.

FIG. 5 shows a filter count and a window size set in each of the horizontal direction evaluation value generation circuit 62 and the vertical direction evaluation value generation circuit 63.

FIG. 6 is a drawing for explaining each window size. FIG. 7 is a drawing for showing weight data W set for each evaluation value E, respectively.

FIG. 8 is a drawing showing a divided area divided by the area search circuit 38.

FIG. 9 is a drawing showing a specific circuit configuration of the area search circuit 38.

FIG. 10 to FIG. 5 are drawings for explaining the processing operation for focusing. FIG. 16 is a diagram for explaining a processing operation for determining a target subject.

FIG. 17 is a drawing for showing the movement of the lens when determining the direction in which the lens is moved to adjust the focus. FIG. 18A and FIG. 18B are drawings showing a state when a non-target object enters the window.

FIG. 19 is a diagram for illustrating a change in the evaluation value stored in the RAM 66 when the moving direction of the lens is determined.

FIG. 20 is a diagram showing data stored in the RAM 66 during an autofocus operation.

FIG. 21 is a drawing for showing changes in the evaluation value obtained by the autofocus operation.

FIG. 22 is a drawing for showing an imaging state of an object B and an object B having the same color as the target object.

FIG. 23 is a diagram showing an object information table.

FIG. 24 is a diagram showing a target object history table. Best mode for carrying out the invention

Hereinafter, embodiments of the present invention will be described with reference to FIGS. 1 to 24.

First, the overall configuration of the video camera device of the present invention will be described with reference to FIG. The video camera device has a lens block 1 for optically condensing the incident light on the front of the imaging device, and an imaging device for converting the incident light from the lens block into an RGB electrical imaging signal. Block 2, a signal processing block 3 for performing a predetermined β processing 1 on the imaging signal, a lens block 1, an imaging block 2 and a ^ processing block. CP IJ 4 is provided.

Lens block 1 can be attached to and detached from the video camera Is provided. This lens block 1 is an optical element that moves along the optical axis to continuously change the focal length without changing the position of the image point, thereby zooming the image of the subject. Zoom lens 11 for focusing, a focus lens 12 for focusing on the subject, and the amount of light incident on the front of the image sensor by varying the aperture And an iris mechanism 13 for adjusting the pressure.

The lens block 1 further includes a position detection sensor 11a for detecting the lens position of the zoom lens 11 in the optical axis direction, and a drive for moving the zoom lens 11 in the optical axis direction. Detects the position of the zoom lens drive circuit 11 for providing a drive control signal to the motor 11 and the drive motor 11 b, and the lens position of the focus lens 12 in the optical axis direction. A position detection sensor 12a, a drive motor 12b for moving the focus lens 12 in the optical axis direction, and a focus lens for supplying a drive control signal to the drive motor 12b Lens drive circuit, 12 c, a position detection sensor 13 a for detecting the opening position of the iris mechanism 13, a drive motor 13 b for opening and closing the iris mechanism 13, An iris mechanism drive circuit 13c for providing a drive control signal to the drive motor 13b is provided. I have.

The detection signals of the position detection sensors 11a, 12a and 13a are always

The control signal from the CPU 4 is supplied to the zoom lens drive circuit 1] (;, the focus lens drive circuit 12 c, and the iris mechanism drive circuit 13 c, while being sent to the CPU 4. It is electrically connected so that

Lens block 1 contains the focal length data and aperture ratio data of zoom lens 1] and the focal length data and aperture ratio data of focus lens 12. It has an EEROM 15 for storing the manufacturer's name and the number of the lens. Each data stored in the PROM 15 is connected to the CPU 4 so as to be read based on a read command from the CPU 4.

In the lens book ^{1, the} direction of the optical axis of the lens is controlled by a pan-Z tilt drive mechanism 16. The control signal from CPU 4 is

When supplied to the motor drive circuit 16b, the motor drive circuit 16b supplies a pulse corresponding to the control signal from the CPU 4 to the drive motor 16a. Thus, the direction of the front axis can be variably controlled by a command from the CPU 4. The pan-tilt driving mechanism 16 is not limited to the one provided in the lens block as described above, but may be provided as a mechanism for driving the entire bidet talent camera. o

The imaging block 2 is a color separation premix for separating the incident light of the lens block into three primary colors of red (R), green (G) and blue (B). The light of the R component, the G component, and the B component separated by the color separation mechanism 21 is imaged on the imaging surface, and the imaging light of each of the formed color components is Image sensors 22 I, 22 G, and 22 B that convert and output the image signals (R), (G), and (B) respectively. For example, the imaging devices 22 R, 22 G, and 22 C i C C D (C a r g e C υ 1 ed D e vi r, o j force).

The image pickup block 21 amplifies the levels of the image pickup signals (R), (G), and (B) output from the image pickup elements 221, 220.22B, respectively, and resets them by Preamplifier for performing censoring to remove set noise 2 3 I? , 23 G, and 23 B.

In addition, the imaging block 2 uses the reference clock from the internally provided reference clock generation circuit to control each of the components in the video camera. Timing signal generation circuit 24 for generating VD signal, HD signal, and CLK signal, which are basic clocks when the circuit operates, and VD signal, HD supplied from timing signal generation circuit A CCD drive circuit 25 for giving a drive clock to the image sensor 22 R, the image sensor 22 G, and the image sensor 22 B based on the signal and the CLK signal is provided. The VD signal is a clock signal representing one vertical period, the HD signal is a clock signal representing one horizontal period, and the CLK signal is a clock signal representing one pixel clock. A timing clock composed of the VD signal, the HD signal, and the CLK signal is supplied to each circuit of the video camera device via the CPU 4, though not shown, although not shown. .

The signal processing block 3 is provided inside the main unit of the video camera, and is a predetermined signal for the imaging signals (R), (G), and (B) supplied from the imaging block 2. Block for processing. The signal processing block 3 converts the imaging signals (R), (G), and (B) from analog to digital video signals (R), (G), and (B), respectively. R, 31G, 31B and a gain control signal for controlling the gain of the digital video signals (R), (G), and (B) based on the gain control signal from the CPU 4. Channel 0 32 R, 32 G, 32 B, digital video signals (R), (G)

, (B) are provided with signal processing circuits 33 R, 33 G, and 33 B for performing predetermined signal processing. The signal processing circuits 33 R, 33 G, 33 B are, for example, two circuits for compressing a video signal at a certain level or more, and a circuit 33 1 R, 33 G, 33 1 B, Compensation circuit that compensates the level according to the set curve. 3 3 2 R, 3 3

2 G, 33 2 B, and black level below predetermined level and white level above expected level ^ B Z W Clip circuit 3 3 3 R, 3 3 3 G,

It has 3 3 3 B. This signal processing circuit 33R, 33G, 33B may include a known black τ correction circuit, contour enhancement circuit, linear matrix circuit, and the like, in addition to the knee circuit, the correction circuit, and the circuit BZW clip circuit.

The signal processing puck 3 receives the video signals (R), (G), and (B) output from the signal processing circuits 33R, 33G, and 33B, and outputs the video signals (R), (G). ), (B), an encoder 37 for generating a luminance signal (Y) and a color difference signal (R-Y) (B-Y).

The signal processing block 3 further receives the video signals (R), (G), and (B) output from the gain control circuits 32R, 32G, and 32B, and outputs the video signals. A force control circuit 34 for generating evaluation data E and direction data Dr for controlling force based on the signals (R), (G) and (B), and signal processing; Receives video signals (R), (G), and (B) output from circuits 33R, 33G, and 33B, and based on the signal levels, image sensors 22R, 22G

Iris control circuit 3 that controls the iris so that the amount of light incident on the 2B and B light is appropriate, and the output from the processing II circuit 3 3R, 3G, and 3B. A white balancer for receiving a video signal (K), (G), and (B), and performing white balance control based on the signal level; and a balance control circuit 3 (;).

The iris control circuit 35 includes an NΛM circuit for selecting a signal having the highest signal level among the supplied video signals (K), (G), and (B), and By dissecting the error in both areas of the signal, the video and σ in each area are fully integrated. The iris control circuit 35 determines all lighting conditions, such as backlighting, normal lighting, flat lighting, spot lighting, etc., of the subject based on the integrated data for each area. Judgment is made, an iris control signal for controlling the iris is generated, and this iris control signal is sent to the CPU 4. C The PU 4 sends a control signal to the iris drive circuit 13c based on the iris control signal.

The white balance control circuit 36 generates a white balance from the supplied video signals (R), (G), and (B) such that (R—Y) = 0 and (Β—Υ) = 0. A control signal is generated, and this white balance control signal is sent to the CPU 4. The CPU 4 supplies a gain control signal to the gain control circuits 32R, 32G, and 32B based on the white balance control signal.

Further, the signal processing block 3 includes an error search circuit 38 and a frame memory 33 ⁽ ).

The relay search circuit 38 removes the luminance signal (Y) from the encoder 37 and the chrominance signals (R-Y) and (B-Y), and based on the luminance signal and the chrominance signal, This is a circuit for selecting, from among the areas set for the entire screen, an area having pixel data that matches the color of the subject specified as the target object. Details will be described later.

The frame memory 39 receives the luminance signal (Y) and the chrominance signals (R-Y) and (B-Y) from the encoder 37, and temporarily stores the luminance signal and the chrominance signal. There are three types of memory for storing: a frame memory for a luminance signal (Y), a frame memory for a color signal (RY) ffl, and a frame memory for a color signal (B-Y). It consists of two frame memories. The luminance signal and color signal stored in each frame memory are read out based on the door address data supplied from the CPU 4, and the read luminance signal and color signal are read out by the CPU 4. Supplied to

Next, the focus control circuit 34 will be described in detail with reference to FIG.

The focus control circuit 34 includes a luminance signal generation circuit Π〗, a horizontal direction evaluation value generation circuit 62, a vertical direction evaluation value generation circuit 63, and a microcomputer. It consists of a black computer 64.

The luminance signal generation circuit 61 is a circuit that generates a luminance signal from the supplied video signals R, G, and B. In order to determine whether the force is correct or not, it is necessary to determine whether the contrast is high or low. Therefore, since the change in contrast is irrelevant to the change in the level of the color signal, by detecting only the change in the level of the luminance signal, whether the contrast is high or low can be determined. You can judge.

The luminance signal generation circuit 61 converts the supplied video signals R, G, B into

By performing a known operation based on Y = 0.3R + 0.59G + 0.11B... (1), the luminance signal Y can be generated.

The horizontal direction evaluation value generation circuit G2 is a circuit for generating a horizontal direction evaluation value. Here, the evaluation value in the horizontal direction is data indicating how much the level of the luminance signal changes when the luminance signal is sampled in the horizontal direction, in other words, how much the horizontal direction This data indicates whether there is trust.

Horizontal evaluation value generating circuit 6 2 generates a first horizontal evaluation value generating circuit 6 2 a for generating a horizontal evaluation value E, the笫〗, the second horizontal evaluation value E ₂ A second horizontal evaluation value generation circuit 62 for generating a _third horizontal evaluation value E ₃ and a fourth horizontal evaluation value circuit for generating a _third horizontal evaluation value E ₃ A _fourth horizontal evaluation value generation circuit 62d for generating E4, a fifth horizontal evaluation value generation circuit 62e for generating a _fifth horizontal evaluation value E5, A sixth horizontal evaluation value generating circuit 6 2 f for generating a _sixth horizontal evaluation value E ₆ and a seventh horizontal evaluation value generating circuit for generating a seventh horizontal evaluation i E 7 6 2 g and the second Eighth horizontal evaluation value generation circuit 6 2 for generating _eight horizontal evaluation values E ₈ and ninth horizontal evaluation value generation circuit 6 for generating a ninth horizontal evaluation value E ₃ 2 i and the 10th horizontal evaluation value E,. A first horizontal direction evaluation value generating circuit 62 j for generating the first horizontal direction evaluation value E, a first horizontal direction evaluation value generating circuit 62 k for generating the first horizontal direction evaluation value E, And a second horizontal direction evaluation value generation circuit 6 21 for generating 12 horizontal direction evaluation values E ₁₂ .

Next, the detailed configuration of the horizontal direction evaluation value generation circuit 62 will be described with reference to FIG.

The horizontal direction evaluation value generation circuit 62a of the horizontal direction evaluation value generation circuit 62, 1 has a high-pass filter 621, which extracts high frequency components of the luminance signal, and an absolute value of the extracted high frequency component. Therefore, the absolute value conversion circuit 622 which makes data all have a positive value, and the data of the high frequency component in the horizontal direction are cumulatively added by integrating the absolute value data in the horizontal direction. The horizontal and vertical integration circuits 6 2 3, the vertical integration circuit 6 2 that integrates the data integrated in the horizontal direction in the vertical direction, the horizontal integration circuit 6 2 3, and the On the other hand, it has a window pulse generating circuit 62 for outputting an enable signal enabling integration operation.

The filter 62 1 filters the high-frequency component of the luminance signal in response to one sample clock CLK from the path 62, which is a 2 pulse generator. It is composed of dance filters. This highpass filter ΰ 2 1

(1-Ζ- ¹ ) / (1-α Ζ-') It also has the cutoff frequency characteristic shown by (2). In the horizontal direction evaluation value generation circuit 62a of ^ 1, << 0.5 is set, and a frequency characteristic corresponding to this "is shown. The window pulse generating circuit 625 includes a plurality of clocks operating based on VD representing one vertical period, HD representing one horizontal period, and CLK representing one sample clock supplied from the CPU 4. Have The window pulse generation circuit G 25 supplies an enable signal to the horizontal integration circuit 62 3 every one sample clock CLK based on the count value of the counter. The enable signal is supplied to the vertical integration circuit 6 2 4 in one horizontal period. In the window pulse generation circuit 62 of the first horizontal evaluation value path 62 a, the counter size is set so that the size of the window becomes 92 pixels × 60 pixels. Initial counter ¾ is set. Therefore, the horizontal direction evaluation value E, output from the horizontal direction evaluation value generation circuit 62, is the data obtained by integrating all the high-frequency components existing in the window of 192-pixel X 60-pixel. It shows that. The counter is connected so that an offset value is supplied from the CPU 4. In the initial count i, a count value is set so that the center of the window coincides with the center of the imaging screen. In contrast, the OH Fuse' bets value supplied from the CP to ¹ 4, it is meant mosquito window down bets值to be added to the initial Ca window down bets value. Therefore, when the offset value is supplied from the CPU 4, the counter value of the counter is changed, and accordingly, the center position of the window is changed.

The other second horizontal evaluation value generation circuits 6 2 b to 12 th horizontal evaluation value generation circuits 62 1 h are similar to the above-described ^ 1 horizontal evaluation value generation circuit 62 a, and are high pass filters. A horizontal integration circuit 623, a vertical integration circuit 624, and a window pulse generation circuit 625. What is different in each circuit is that the filter coefficients set in each circuit (62a to 621) and the window Combinations of sizes are different.

Therefore, the 12 evaluation values E ₁ ,..., E 12 generated by the respective circuits have different values.

In FIG. 5A, the first horizontal evaluation value generation circuit 62 a to the 12th horizontal evaluation value generation circuit 62 1 determine what value of the filter coefficient and the window size are set. The reason why such different filter coefficients are set is described below.

For example, a hi-no filter with a high cut-off frequency is very suitable near just focus (meaning that the focus is on). This is because the evaluation value has a large change rate with respect to the movement of the lens near the just focus. Also, even if the lens is moved where the focus is greatly deviated, the rate of change of the evaluation value is small, so a high-pass filter with a high cut-off frequency has a low focus. It is not suitable where there is a large gap.

On the other hand, a high-pass filter with a low cut-off frequency is suitable where the power is significantly shifted. This is because when the lens is moved where the focus is largely shifted, the rate of change of the evaluation value is large. Even if the lens is moved in the vicinity of the just-force, the change rate of the evaluation value is small, so that the high-pass filter with a low cut-off frequency is not close to the just-focus. However, it is not suitable. In other words, both high-pass filters with a high cut-off frequency and high-pass filters with a high cut-off frequency have advantages and disadvantages, and which filter is the most suitable. Cannot be said in one word. Therefore, it is more desirable to generate a plurality of evaluation values using a high-pass filter of a plurality of filter coefficients and select the most appropriate evaluation value. In the horizontal direction evaluation value generation circuit 63 in this embodiment, a plurality of types of windows are set as shown in FIG. 6A. If no offset value is supplied from the CPU 4 to the counter inside the window pulse generation circuit 6 25, the center of these multiple windows will be the center of the imaging screen. Matches. Window W

1 is a 192-pixel X60-pixel window, window W2 is a 13-pixel X60-pixel window, and window W3 is a 384-pixel There is a window of X120 pixels, window W4 is a window of 2G4 pixels X120 pixels, and window W5 is 768 pixels of X120 pixels. There is a window, window W6 is

5 4 8 pixels X〗 20 pixels d queen

By setting a plurality of windows in this way, different evaluation values corresponding to the respective windows can be generated. Therefore, regardless of the size of the subject whose focus is to be adjusted, the second horizontal evaluation value generation circuit β 2 a

An appropriate evaluation value can be obtained from any one of the horizontal evaluation value generation circuits 6 2 1.

Next, the configuration of the vertical direction evaluation generation circuit 63 will be described with reference to FIG. 2 and FIG.

The vertical direction evaluation value generation circuit 63 is a circuit for generating a vertical direction evaluation value. Here, the evaluation value in the vertical direction is data indicating how much the level of the luminance signal changes when the luminance signal is sampled in the vertical direction, in other words, how much the vertical control value is. This data indicates whether there is trust.

The vertical evaluation value generating circuit 6 3, a first vertical evaluation value generating circuit 6 3 a for generating a first vertical ΪΙ direction evaluation value E, _3, a second vertical evaluation value E, ₄ The second to produce! 6 A vertical direction evaluation value generation circuit 6 3 b and a third vertical direction evaluation value E,> A vertical evaluation value generation circuit _63d , a fourth vertical evaluation value generation circuit _63d for generating a fourth vertical evaluation value Elc, and a fifth vertical evaluation value E, _A fifth vertical evaluation value generating circuit 63 e for generating the _seventh vertical evaluation value E 6 and a sixth vertical evaluation value generating circuit 63 f for generating the sixth vertical evaluation values E and _β ; Seventh vertical evaluation value E

_{1 3} and vertical evaluation value generating circuit 6 3 g of a Ί for generating a vertical evaluation value of the 8 E _2. And a ninth vertical evaluation value generating circuit 63 i for generating an ninth vertical evaluation value E ₂ , and an eighth vertical evaluation value generating circuit 63 h for generating Vertical evaluation of 0

J 1

First 1 vertical evaluation value generating circuit for generating a vertical evaluation value generating circuit 6 3 j of the first 60, the first 1 in the vertical direction evaluation value E _{2 3} for generating the value E _{2 2} and 6 3 k, and the first and second vertical evaluation value E _{2 4} and a vertical evaluation value generating circuit 6 3 1 of the first 2 to generate.

Next, referring to FIG. 4, a vertical evaluation value generation circuit

6 Detailed configuration of 3 will be described.

The first vertical evaluation value generation circuit 63 a of the vertical evaluation value generation circuit 63 is a horizontal average value generation circuit 631 that generates -average value data of the level of the horizontal luminance signal, A high-pass filter 632 for extracting the high-frequency component of the average value data of the luminance signal, and an absolute value conversion circuit for converting the extracted high-frequency component to an absolute value, so that all the data have a stop value. 6 3 3, a vertical integration circuit 6 3 4 that integrates the absolute value data in the vertical direction to accumulate and add high frequency component data in the vertical direction, and a horizontal average value generation circuit 6 3 1 And a window pulse generation circuit (] 35) for sending an enable signal enabling integration operation to the vertical integration circuit 63 A.

The high-pass filter 32 is a window pulse regeneration circuit G 25 It consists of a one-dimensional finite impulse response filter that filters the high-frequency component of the luminance signal in response to one horizontal period signal HD from. This high-pass filter 632 has the same cutoff frequency as the high-pass filter 62 1 of the first horizontal direction evaluation value generation circuit 62a. This first vertical evaluation value generation circuit

In 6 3a, "= 0.5 is set, and the frequency characteristic according to this is shown.

The window pulse generation circuit 635 operates based on the VD signal supplied from the CPU 4 representing one vertical period, HD representing one horizontal period, and CLK representing one sample clock. It has multiple counters. This Windo ,. The zero generation fi35 is one sample clock for the horizontal average value generation circuit 631, based on the count value of the count value. A signal is supplied, and an enable signal is supplied to the vertical integration circuit 634 every one horizontal period. In the first vertical direction evaluation value path 63 a, the window pulse generating circuit 63 5 has the initial power of the printer so that the size of the window is 120 pixels x 80 pixels.ゥ¾: is set. Therefore, the evaluation value E, ₃ output from the vertical evaluation value generation path f; indicates the data obtained by integrating the high frequency component in the vertical ifi direction in the window of 120 pixels X 80 pixels. Will be. This counter is connected so that the offset 4 is supplied from the CPU 4. The initial count value is set so that the center of the window coincides with the center of both sides of the imaging. On the other hand, the offset value supplied from the CPU 4 neglected the count ¾: added to the initial count value. Therefore, C

When the offset is supplied from PU 4, the count down of the counter of the window pulse generating circuit 635 is changed, and the center position of the window is changed accordingly. You. The other second vertical evaluation value generation circuits 63 b to the 12th vertical evaluation value generation circuits 63 3 h are similar to the first vertical evaluation value generation circuits 63 a described above. A horizontal average value generation circuit 631, a high-pass filter 632, an absolute value conversion circuit 633, a vertical integration circuit 634, and a window pulse generation circuit 6353. have. What is different in each circuit is that, as in the horizontal direction evaluation value generation circuit 62, the combination of the filter coefficient and window size set in each circuit is different. Has become.

Therefore, 評価 two evaluation values E, to be generated by each circuit

The values of E ₁₂ will have different values.

In Figure 5β, what value of filter coefficient and window size are set for the first vertical evaluation generation circuit G2a to the 12th horizontal evaluation value generation circuit 6 21 Is shown.

In the vertical direction evaluation value generation circuit 63 in the present embodiment,

6 Several types of windows are set as shown in Fig. B. When the offset value is not supplied from the CPU 4 to the power counter inside the window pulse generating circuit 6 25, the centers of these multiple windows coincide with the center of the imaging screen. I do. Window W7 has a window of 120 pixels X 80 pixels, window W8 has a window of 120 pixels X 60 pixels, and window W9 has 24 windows. There is a window of 0 pixel X 16 0 pixels, window W 10 is a window of 240 pixels X 120 pixels, and window W 1 1 is 480 pixels X 3 There is a window of 20 pixels, and window W12 has a window of 480 pixels X 240 pixels.

As described above, by providing a circuit in which a plurality of filter characteristics and a plurality of windows are set, respectively, it is possible to generate different evaluation values corresponding to each combination of each filter coefficient and each window. Can be. Therefore, regardless of the shooting condition of the subject to be focused, a comprehensive evaluation value is generated from multiple evaluation values, and any one of the evaluation values is inappropriate. Even so, an accurate overall evaluation value can be obtained.

Therefore, in the present embodiment, since there are 24 evaluation value generation circuits in which 12 types of window sizes and 2 types of filter coefficients are combined, a plurality of types of evaluation values can be obtained. Further, since an overall evaluation value is obtained based on each evaluation value, the accuracy of the evaluation value can be improved.

1

Next, referring to FIGS. 2 and 9, 9

4 will be described.

Microphone loco computer 6 4, E generated in the horizontal direction evaluation value generating circuit 6 2及beauty vertical evaluation value generating circuit 6 3, both receive 2 four evaluation values of ~ E _{2 4,} the 2 4 Based on the evaluation values, the direction in which the lens moves and the lens position where the evaluation value is maximum

That is, a circuit for determining the in-focus lens position.

The micro computer 64 has R 0 065 storing a program for calculating 24 evaluation values according to a predetermined flow. As shown in FIG. 7, this R〇Μ65 is output from each of the 24 evaluation value generation circuits (62a to 62) and 63a to 63 3, respectively. Twenty-four weight data W i are stored so as to correspond to the twenty-four evaluation values E i (i-1, 2, 2,..., 24). The weight data w; is data for giving a priority to the 24 evaluations E; The higher the value of the weight data W i, the higher the priority of the corresponding evaluation value E; This means that there is high quality. Note that the weight data W i is a fixed value set in advance at the time of factory shipment. The microcomputer 64 has 24 evaluation value generation circuits (62 a to 62 and 63 a to 631) each supplied with 24 power supply circuits. There is a RAM 66 for storing the evaluation value E i (i = 1, 2,..., 24) in association with the position of the focus lens. Here, the evaluation value when the lens is at the position of X,

Let E, (X,) to E ₂₄ (X ₁ ). First, the evaluation values Ε, (X,) to E ₂₄ (X,) generated when the lens is at X, are stored in the RAM 66. Et al is, when the lenses are moved to the position of X ₂ from X, the position, the generated evaluated values when the lenses are moved in the X ₂ E, is _{_{(X 2) ~ Ε 24 (}} X 2) R It is stored in AM66.

RAM 6 6, since de one data is stored in the re-Nguba' off § method, as long as the stored data is not filled, the previously stored evaluation value E, (X,) ~ E 24 (X,) is It will not be erased. Further, these evaluation values E i are stored in the RAM 64 by specifying a pointer by the micro computer G 4. Next, the engineering search circuit 38 will be described with reference to FIGS. 8 to 10.

The error search circuit 38 divides the image capturing area into 12 areas, and the pixel data having the same color as the object set as the target object is output to which division error. It is a circuit to search for existence

. In the area search circuit 38, a determination process is performed by a logic circuit in the area detection circuit 38 on all pixel data supplied from the encoder. Specifically, as shown in Fig. 8, if one G) area is defined to be 481 elements X 30 丽 i elements, it is divided into 16 elements in the horizontal direction, and ¾ Since it is divided into eight in the ifi direction, 128 areas are eventually determined. As shown in FIG. 8, these 128 areas are defined in the application with the area numbers Λ _n (,. To A, ₂₇ . First, a specific configuration of the error search circuit 38 will be described with reference to FIG. First, the luminance signal Y, the color signal i R — ΥI and the color signal IB — YI are supplied from the encoder 37 to the error search circuit 38 for each pixel data. Also, the upper search signal IY from the CPU 4 is sent to the relay search circuit 38. And the lower limit luminance signal i Y

. And the upper limit color signal i R. One Y. And the lower color signal IR. ― Yo and the upper limit color signal IΒ. One Υ. And the lower color signal IB. One Y. IL is supplied. Et al is, encoder 3 7 multiplication factor is multiplied to the supplied luminance signal Y from the _"3, coefficient _4, the engagement number" _5, coefficient _"6 is supplied from the CPU 4.

Here, the upper limit luminance signal 1 Y supplied from the CPU 4. And the lower limit luminance signal I Υ. And the upper color signal IR _D — Y. iu and the lower color signal IR o — Υ. And the upper limit color signal I Β. Υ »and the lower limit color signal I Β. One Y. IL and will be described. First, the luminance signal YD and the color signal IR. One Υ. Color signal IB. One Y _D I is a signal subject force main Raman has set aimed object is obtained on the basis of the luminance signal and color signals冇. Therefore, once the target object is set, this value will not be changed. The luminance signal of a certain pixel data is an upper luminance signal i Y _U and a lower luminance signal IY. If the value is between I and., The upper limit luminance signal ¾ γ ₀ I 及 and the lower limit luminance signal IY. It is set so as to have a value close to the luminance signal II of the object rest, and similarly, the color signal IR-YI of a certain pixel data is the upper limit color signal IR. Lower limit color signal IR-Y. If the Ri of I is ίι, the color signal IR of a certain pixel data-ΥI is almost the same level as the color signals IR0 to Υ0I of the set target object as can be determined that the upper limit Iroshin ^ IR. one Upsilon. and lower color signal IR. one Upsilon _[pi Mr., the first object object It is set to have a value close to the color signal IRYDI of 1. If the color signal IB-Yi of a certain pixel data is a value between the upper limit color signal IBo-Yiυ and the lower limit color signal IΒYIL, the color signal iB-YI of the certain pixel data becomes The upper limit color signal IBY _Q I u and the lower limit color signal i BYI are the second color signal i Β Υ i of the target object so that it can be determined that the color signal i BY i of the target object is at substantially the same level as the color signal i BY i. It is set to have a value close to.

The multiplier search circuit 38 includes a multiplication circuit 71 a for multiplying the luminance signal Y supplied from the encoder 37 by a multiplication coefficient <<: ₄ , and a multiplication coefficient “ ₃ ” for the luminance signal Y. A multiplication circuit 7 1 multiplies the luminance signal Y by a multiplication coefficient <<: s A multiplication circuit 7 1 (: and a multiplication circuit multiplies the luminance signal Y by a multiplication coefficient ₅ 7 1 d and the in and. is found to have, Sui tree et re § search circuit 3 8, for selecting one of the multiplication circuit 7 1 th power calculating force from a and upper color signal _IR Υ π I u a latch circuit 7 2 a, multiplication output from the multiplier circuit 7 1 b and lower color signal IR ₀ -. Y 0 I, a switch circuit 7 2 b for selecting either of the multiplication circuit 7 1 c Multiplied output and upper limit color signal i B.-Switch circuit 7 2 c for selecting one of YI と and multiplied output from multiplier circuit 7 1 d and switch circuit for selecting one of lower limit color signal IBY _u ! Switch times In addition, the error detection circuit 38 includes a comparator 73a supplied with the luminance signal Y and the upper limit luminance signal IY, a comparator 73a supplied with Iu, and a luminance signal Y with the lower limit. A switch 73 3c to which a luminance signal IY, IL is supplied, and a comparator 73 C to which a signal from the switch circuit 72A and a color signal | R-YI are supplied. A comparator 73d supplied with the signal from the circuit 72b and the color signal IR-YI, and a comparator 733 supplied with the signal from the switch circuit 72c and the color signal IB-YI c and the signal from the switch circuit 72 d. And a comparator 73 f to which signals and color signals IB-YI are supplied. Further, the area search circuit 38 includes a gate circuit 74 a to which the output of the comparator 73 a and the output of the comparator 73 b are supplied, a con- A gate circuit 74b supplied with the output of c and the output of the comparator and the レ 3 d, and a gate circuit supplied with the output of the comparator 73e and the output of the comparator 73f. And a gate circuit 75 to which the output of the gate circuit 74a, the output of the gate circuit 74b, and the output of the gate circuit 7c are supplied.

Further, the area search circuit 38 has a flag signal generation circuit 76 'composed of 128 chip circuits. 1 2 8 Chi-up circuit, Figure 8 lambda _{D 0} shown in, from lambda, of _{2 7:} is provided so as to correspond to the 1 2 8 E Li §. The output of the gate circuit 75, the pixel clock CLK, and the chip select CS are supplied to each of the chip circuits. The pixel clock CL # and the chip select CS are supplied from the CPU 4 so as to correspond to the luminance signal and the color signal supplied from the encoder 37 to the pixel data ^. The pixel clock CLΚ is a clock corresponding to the processing timing of each pixel data, and is used for processing pixel data and pixel data processed by the preceding logic circuit. If they match, a “Low” signal is provided to the chip circuit, and at other times

Provides a "High" signal. In the chip select CS, the “Low” signal is supplied only to the selected chip circuit out of the 128 chip circuits, and the “High” signal is supplied to the other unselected chip circuits. .

Each chip 0 path provided in the flag signal generation circuit 7G has a gate circuit 76a and a counter 6b, respectively. Therefore, the flag signal generation circuit 76 has 128 gate circuits 7 (; a and 128 counters 76 b. This gate circuit 76 a , Supply The gate circuit 75 outputs “Low” only when all of the output of the gate circuit 75, the pixel clock CLK and the chip select CS are “Low”. The counter 76b is a counter that counts up only when “Low” is supplied from the gate circuit 76a in response to the clock timing of the pixel clock CLK. A flag signal is generated when the count value exceeds a predetermined number (5 or more in this embodiment). The generated flag signal is supplied to a multiplexer 77.

The multiplexer 77 receives the flag signal output from each chip circuit of the flag generation circuit ΊG, and supplies the signal to the CPU 4. At this time, the multiplexer 77 sends to the CPU 4 a signal of the chip circuit from which the flag signal has been output. The CPU 4 can select an area in which pixel data having the same color component as the target object exists based on the examination.

First, in the area search circuit 38, the switch circuits 72a, 72b.72c and 7c provided in the area search circuit 38 before the area search processing is performed. Since the 2d switching operation must be performed, this switching operation will be described.

First, the switch circuits 72a, 72b,? In order to switch 2c and 72d, the luminance signal I Y of the object set by Cameraman as the target object. I, color number I R «-Y n! And signal IB. — Y. It is necessary to select the subject mode based on I. In this subject mode, four modes are set, and mode 0 to mode 3 will be sequentially described below.

Mode 0 is the mode selected when the subject set as an EHl object has some color information. That is, i R representing the color component of the object. -Y n I 及び and I Y Y. I Ζ Υ means that 0 is above a certain level. Then, when the color of the selected target object is strong u

The CPU 4 sets the luminance signal IY of the set object rest. I, color signal i R. One Υ. I and color signal i B. The relationship of one Y _n I is α 3 XIY 0 R 0-Y x Y

And

a 5 XIY. ≤ | Β. -Y. | ≤ " _c x | Y. |

-Mode 0 is selected when the condition indicated by (70) is included. When the mode () is selected as the subject mode by the CPU 4, the CPU supplies a control signal to the switch circuits 72a, 72b, 72c and 72d. The switching states of the switch circuits 72a, 72b, 72c and 72d are "1JP", "UP", "UP" and "UP", respectively. Once switched, this switching state does not change until the subject mode is changed.

Mode 1 is a mode selected when the color component of the subject set as the zero-like object has a red component equal to or higher than a predetermined level but does not have a color component equal to or higher than the predetermined level. In other words, IR that represents the color component of the subject. One Y. ΖΥ ΖΥη is above a certain level, but I Β. One Υ. | ΖΥ. Is not higher than a predetermined level; Therefore, the CPU 4 sets the luminance signal | Y _Q I and the color signal IR of the set target object. The relationship of one Y _D I is

« ₃ x | Y. i ≤ | R. -Y. | α · ₄ x IY _n I and

IB. One Y. i ≤ « ₅ x | Y. |

Mode 1 is selected when the conditions shown in (71) are included. CPU 4 When the mode 1 is selected as the subject mode, the CPU 4 supplies a control signal to the switch circuits 72a, 72b, 72c and Ί2d, and The switching states of the circuits 72a, 72b, 72c and 72d are "UP", "UP- |", "D〇WN" and "D〇WN", respectively.

Mode 2 is a mode selected when the blue component of the color component of the subject set as the target object is higher than the predetermined level but the red component is not higher than the predetermined level. In other words, IB that represents the color component of the subject. One Y. IXYO is above a certain level, but i R. One Υ. | It means that _{Π Π} is not above the specified level. Therefore, the CPU 4 sets the luminance signal Y of the set target object. , The relationship of the color signal IB o-Y o I is

R Y x Υ 0

And

x Υ 0 Β ο-Υ ο a χ I Υ π I

(72)

Mode 2 is selected when the conditions shown by are included. When mode 2 is selected as the object mode by CPU 4, CPU 4 supplies a control signal to switch circuits 72a, 72b, 72c and 72d. The switching states of the switch circuits 72a, 72b. 72c and Ί2d are set to "DOWN", "D〇WN", "UP" and "UP", respectively.

Mode 3 is a mode selected when both the red component and the blue component of the color component of the subject set as the target object do not exceed a predetermined level. In other words, IR that represents the color component of the subject. One Y. | 丫. And | — 丫. I Ζ Υ. , Both mean that there is no more than a predetermined level. In other words, this setting It means that the target object has no color. Therefore, the CPU 4 outputs the luminance signal Y _D and the color signal IR of the set target object. ― Y 0 I and color signal IB. One Y. Mode 3 is selected when the relation of I does not apply to the above equations (70), (71) and (72). When the mode 3 is selected as the subject mode by the CPU 4, the CPU 4 supplies a control signal to the switch circuits 72a, 72b, 72c and 72d, and The switching states of the switch circuits Ί 2 a, 72 b. 72 c and 72 d are referred to as “D OWN”, “D〇WN”, “D OWN” and “D OWN”, respectively. You.

When the above-described switching operation is completed, the area search circuit 38 performs an object search processing operation. Next, this search processing operation will be described in order so as to correspond to each subject mode with reference to FIG.

1) When mode 0 is selected

First, regardless of the selected mode, the comparator 73a outputs the upper limit luminance signal IY. By comparing I u with the luminance signal Y supplied from the encoder 37,

Y ≤ i Y. I

In the case of, “H i g h” is output,

I Yo I u ≤ Y

In the case of, “Low” is output. Also, the comparator 73 b is

, Lower luminance signal I Y. I L and the luminance signal supplied from the encoder

Υ comparison,

I Υ. ≤ Υ

In the case of, "II i g h" is output, and

Y ≤ I Y.

In the case of, “Low” is output. The gate circuit 74a is When the output signal from the comparator 73 a and the comparator Ί 3 b is received, and both the output signals from the comparators 73 a and 73 b are “High”, “Lo” wj is output to the subsequent gate circuit 75.

That is, the operations performed by the comparators 73a and 73b and the gate circuit 74a are as follows.

I Yo I L ≤ Υ ≤ 卜. (74). In other words, if the luminance signal から supplied from the encoder is included in the condition defined by the equation (74), “Low” is output from the gate circuit 74a. is there.

Next, when mode 0 is selected, the switching states of the switch circuits # 2a and # 72b are "UP" and "UP", respectively. The comparator 7 3 c is supplied with YX ₄ and i R—Y i, and the comparator 7 3 d receives YX " ₃ and IK-Y

I will be supplied. The luminance signal Y and the color signal IR-YI are data supplied from the encoder 37. Comparator 73c compares YX " ₄ with IK-YI,

IR-Y '≤ YX a ₄

In the case of, "g I i g h" is output,

YX ₄ ≤ IR-YI

In the case of, “L 0 w” is output. Further, the comparator 73d compares Yx " ₃ with IK-YI,

Y X α 3 ≤ I R-Υ I

In the case of, "il i g h" is output, and

IR-Y: ≤ YX _a

In the case of, “L 0 w” is output. The gate circuit 74b receives the output signals from the comparator 73c and the comparator 73d, and When the output signals from the comparators 73c and 73d are both "High", "L0w" is output to the gate circuit 75 at the subsequent stage.

That is, the operation performed by the comparators 73c and 73d and the gate circuit 74b is

Y X α 3 ≤ I R-Υ I ≤ Υ X a... (75) That is, the color signal supplied from the encoder

If 1R—ΥI is included in the condition defined by this equation (75), “Low” is output from the gate circuit 74b.

If mode 0 is selected, the switch circuit 7

The switching states of 2c and 72d are "UP" and "UP", respectively. YX " _G and IB-YI are supplied to the comparator 73e, and YX" ₅ and IB-YI are supplied to the comparator 73I ". Note that the luminance signal Y and the color signal I — Υ I

And data supplied from the encoder 37. Comparator Ί

3 e performs the comparison between YX α ' ₆ and I Β— Υ |

I Β-Υ I ≤ Υ X α β

In the case of, "II〗 g h" is output, and

Ύ X _r , ≤ | Β-Υ |

In the case of, “Low” is output. Also, the comparator 73 f is

, Y x " ₅ and i B_Y I

Ύ X ₅ ≤ IB-YI

In the case of, “H i g h” is output,

I B-Y i ≤ Y x a r,

In the case of, “Low” is output. The gate circuit 74c receives the output signals from the comparator Ί3e and the comparator Ί3f, and outputs the signals HHig If “h”, “Low” is output to the gate circuit 75 at the subsequent stage.

That is, the operations performed by the comparators 73e and 73f and the gate circuit 74c are as follows.

Y X α 5 ≤ I Β-Υ I ≤ Ύ X a e--· · (7 6) In other words, if the color signal i BYI supplied from the encoder is included in the condition defined by the equation (76), “Low” is output from the gate circuit 74c. And

The gate circuit 75 receives the output signals from the gates 74a, 74b, and 74c, and outputs all the signals from the gates 74a, 74b, and 74c. Only when “High”, “L 0 w” is supplied to each chip circuit of the flag generation circuit 76.

That is, the mode 0 is selected as the subject mode when the conditions of the equations (70a), (70b) and (70c) are satisfied. The calculations performed by the gate circuit 74a-73c, the gate circuit 74a-73c, the gate circuit 74a, and the gate circuit 75

IY o I _L Υ Υ

And

YX a ₃ ≤ R Υ ≤ Ύ X a

And

Y X α 5 ≤ Β— Υ Ύ X a

(700). When the mode 0 is selected, the fact that the condition of the equation (700) is satisfied means that the luminance signal ¾Y and the color signal II of the pixel data supplied from the encoder 37? I 色 I and the color signal i Β— Υ I are the brightness of the subject set as the target object Signal Y. , Color signal IR. One Y. I & color signal IR. ― It means that it almost matches YDI. Then, only when the color of the target object and the color of the pixel data match, “L 0 w” is output from the gate circuit 75.

2) When mode 1 is selected

When mode 1 is selected, the operation is exactly the same as when mode 0 is selected, so a detailed description is omitted.

Since the operation is the same as that when mode 0 is selected, when mode 1 is selected as the subject mode, comparators 73a to 73f and gate circuit 74a ~ 74c and the operation performed by the gate circuit 75 are

I Y. Y Y 0

And

Ύ XRY ≤ Ύ X a ₄

And

! B 0 Y. I L ≤ I B-Y I ≤! B 0 Y.

•… When mode 1 that can be defined by (701) is selected, the fact that the condition of this (701) expression is satisfied means that the pixel data supplied from the encoder 37 is satisfied. The luminance signal Y, the color signal IR-YI and the color signal IB-Yi are the luminance signal Y of the subject set as the target object. , Color signal IR. One Y. And color signal i R. -Υ. It means that it almost matches I. Then, as in the case of the mode 0, the gate circuit 75 outputs “Low” only when the color of the target object and the color of the pixel data match.

3) When mode 2 is selected

When Mode 2 is selected, Mode 0 and Mode Since the operation is the same as when node 1 is selected, detailed description is omitted.

Since the operation is the same as when mode 0 is selected, when mode 2 is selected as the subject mode, the comparators 73 a to 73 f and the gate circuit 74 a To 74 c and the operation performed by the gate circuit 75 are:

I Y. Y Y 0

And

I R 0-Y 0 R— Y R n-Y 0

And

YX or 5 ≤ B-Y ≤ Ύ a _B

· · · · (702) When the mode 2 is selected, the fact that the condition of the expression (702) is satisfied means that the luminance signal Y, the color signal IR—ΥI, and the color signal IR of the pixel data supplied from the encoder 37 are satisfied. Color signal i B — YI force Brightness signal Y of the subject set as the target object. , The color signal i R. One Y. I and color signal IR. One Y. It means that it almost matches I. As in the case of the mode 0, only when the color of the target object and the color of the pixel data match, "Low" is output from the gate circuit 75.

4) When mode 3 is selected

The operation when mode 3 is selected is exactly the same as the operation when mode ϋ, mode 1 and mode 2 are selected, and a detailed description thereof will be omitted.

Since the operation is the same as when mode 0 is selected, when mode 3 is selected as the abort mode, comparators 73a to 73f, gate Circuits 74a-74c and gate circuit 75 The operation performed is

I Y. Υ Υ 0

And

Y X α 3 ≤ I R-Υ Υ X α 4

And

Υ X α 5 ≤ I Β-Y I ≤ Ύ x a

(703). When the mode 3 is selected, the fact that the condition of the expression (703) is satisfied means that the luminance signal Y, the color signal IR—ΥI, and the luminance signal Y of the pixel data supplied from the encoder 37 are satisfied. Color signal I Β— The luminance signal 被写体 of the subject whose I is set as the target object. , Color signal IR. One Υ. I and color signal IR. One Υ. It means that it almost matches I. Then, as in the case of the mode 0, the gate circuit 75 outputs “Low” only when the color of the target object and the color of the pixel data match.

Further, referring to FIG. 9, mode 0 is selected as the subject mode, and a plurality of elementary data having the same color as the object rest are displayed. of e Li Ryo lambda _3:, Oh, an example in which only the Gaité, describing the overall勁作of e re a search circuit 3 8.

First, a luminance signal Υ, a chrominance signal IR-ΥI, and a chrominance signal IΒ—Yi are sequentially supplied from the encoder 37 to the encoder search circuit 38 so as to correspond to the raster scan. . In other words, all pixel data from the encoder 37 is supplied to the relay search circuit 38, and it is determined whether or not each pixel data is included in the condition of the equation (700). Done. Note that all pixel data is supplied to the area search circuit 38, but the determination as to whether or not the data is included in the condition defined by the equation (700) is performed by the switch circuit. Depending on the hardware circuits such as 72 a to 72 d, cono, ° recorder 73 a to 73 f, and gate circuit 74 a to 74 c Since the processing is performed, the judgment processing can be performed in real time without putting a processing load on the CPU 4.

In this example, the addition et rear A _{0 3 5,} since there is no pixel data having the same color as the set purpose objects, et Li A lambda. _Even if any pixel data existing in the area other than the area 35 is supplied to the area search circuit 38, the gate circuit 75 outputs "I igh". When pixel data having the same color as the target object in the area A _{0 35} is supplied to the area search circuit 38, the gate circuit 75 first outputs “L 0 w

Output J. At this time, the chip select CS supplies “Lowj” only to the 36th chip circuit corresponding to the area No. 03, and supplies “Highg to the other chip circuits. The pixel clock supplies “Low” to the chip circuit at a timing at which pixel data having the same color as the S-like object is supplied. Therefore, the gate circuit 76'a in the 36th chip circuit. _3-5, gate circuit 7 Five et gamma L ow "is supplied, as a pixel click locked CLK is supplied" L 0 w ", and the chip select is supplied" L ow " Only at this time, the gate circuit 76a. ₃ 5 sets “Low” to the counter 7

6 boar, supplied to Counter 76 b. _{3 5} counts up when “Low” is supplied from the gate circuit 7 G a 0 3 5. When this count value becomes 5 counts or more, the flag signal is multiplexed.

7 Output to 7. What does this mean? _In 35, there are 144 0 pixel data, and if there are 5 or more pixel data that match the color of the set target object, the flag signal is output to the area 5. . _{3 5} is that the output chip circuit or et corresponds to. Therefore, in this example, the power of the chip circuit corresponding to the area 0 35 is 7 fib. ₃ to ₅ only outputs the flag signal, Do flag signal from Chikarau printer 7 6 b of the chip circuit corresponding to the d rear lambda _{0 3 5} non-e re A A _{[pi 3 5} is outputted No.

The multiplexer 77 outputs the plug signal output from each chip circuit to the CPU 4 so as to correspond to the end of the plug signal. In this case, it supplies the flag signal is output from the 3 sixth chip circuit corresponding to the e Li A AD _{3 5} to CPU 4.

The operation of the end search circuit 38 composed of hardware in this way allows the CPU 4 to determine in which of the pixels the pixel having the same color as the set target pixel exists. Can be recognized in real time. Next, the operation of the video camera device will be described with reference to the flowchart of FIG. 10 to FIG.

First, the force control operation will be described.

The transition from manual focus to auto focus can be performed by pressing the auto focus button provided on the operation unit 5 by the cameraman. The focus mode is set. Autofocus mode is a continuous mode that, once pressed, continues in autofocus mode until commanded to switch to manual focus. The auto-focus mode is stopped when the focus is adjusted, and the non-continuous mode in which β automatically shifts to the manual mode. The following description of the flowchart is for the continuous mode. In steps S100 to S131, a process for determining in which direction the focusing lens is to be moved is performed. In steps S201 to S221, Processing is performed to determine the lens position at which the evaluation value is maximized.

First, in steps S100 to S104, as shown in FIG. 17, the focus lens is moved to the lens initial position X ( , At a distance of D / 2 in the Far direction Go to position X, that, X, from the N ear direction moves to position X ₂ at a distance and D, i.e. the position from chi ₂ to the distance DZ 2 in F ar direction, the lens initial position X _D Perform a return operation. The Near direction means a direction approaching the image sensor, and the Far direction means a direction away from the image sensor. D represents the depth of focus. Microcomputer 64 has lens position X. , X,, X 2, the evaluation value E i (X.), the evaluation ifiE i (X evaluation value E i () generated by the horizontal evaluation value generation circuit 62 and the vertical evaluation value generation circuit 63. X 2) is stored in RAM 6.

In here, to explain the internal medicine move DZ 2 more than a full O carcass lens Why from X _Q. Depth of focus is data that indicates the range where the focus is at the center of focus. Therefore, even if the focus lens is moved within the depth of focus, the shift of the focus cannot be recognized by human eyes. Conversely, the lenses X, when moving to a pressurized et X _2, is moved over the focal depth, off O over mosquito scan of deviation that will appear in the image pickup signal by the mobile. In other words, by setting the maximum movement of the lens within the depth of focus, the displacement of the force cannot be recognized.

Next, each of the steps S100 to S104 will be described with reference to FIG. 17 for details.

In Step S 1 0 0, microphones π co down computer 6 '4, horizontal direction toward evaluation value generating circuit 6 2 and the evaluation value generated in the vertical direction evaluation value generating circuit 6 Ε, (Χ ₀₎ ~ evaluation value Ε stores ₂₄ (chi _0), the. When the storage is completed, the microphone π-computer G 4 instructs the CPU 4 to move the focus lens by a distance of DZ 2 in the Far direction.

In step S101, the CPU 4 outputs a command to the focus lens motor drive circuit 12c, and outputs the focus lens. Move in the Far direction by a distance of D2.

In step S 102, the micro computer 64 receives the evaluation values Ε, (,) newly generated in the horizontal evaluation value generation circuit 62 and the vertical evaluation value generation circuit 63. X,) to the evaluation value Ε ₂₄ (X,) are stored in RAM 66. After storing the evaluation value, the micro computer 64 issues a command to the CPU 4 to move the focus lens by the distance D in the Near direction.

In step S103, the CP IJ 4 outputs a command to the focus lens motor drive circuit 12c to move the focus lens by the distance D in the Near direction. Move to

In step S104, the microphone π computer 64 outputs the evaluation values E, (), which are newly generated in the horizontal evaluation value generation circuit 62 and the vertical evaluation value generation circuit 63. X ₂ ) to evaluation value E ₂₄ (X ₂ ) are stored in RAM 66. When the storage is completed, the microphone π computer 64 instructs the CPU 4 to move the focus lens by a distance of D / 2 in the Far direction.

Therefore, when step S104 is completed, the lens position X is stored in the RAM 66 of the micro computer 64. The evaluation values E, (X ₀ ) to E ₂₄ (X.) at, the lens position X, and the evaluation values E, (X) to E 2 (Χ) at lens position X.評価, (Χ ₂ ) to Ε ₂₄ (X

2) is recorded.

Steps S105 to S115 are tips for selecting an inappropriate evaluation value from the 24 evaluation values.

First, with reference to FIG. 18A and FIG. 18B, the operation of step S105 to step S115 will be described in a comprehensive manner.

FIGS. 18A and 18B show an image of a target object A whose window force is to be adjusted within window W2, which is outside window W2. Inside window W1, target object A This shows a state in which the non-target object B having a high contrast existing in front of the object is being imaged. At this time, since the object B exists in the frame of the window ゥ W 1, the first horizontal direction evaluation value generation circuit 62 a in which the value of the window W 1 is set is obtained. The generated evaluation value E, is combined with the high-frequency component of the object B, and is not appropriate as the evaluation value of the object A. Therefore, the value of the evaluation value E, is considerably larger than the value of the evaluation value E ₂ generated by the second horizontal evaluation furnace generation circuit G 2 b in which the value of the window W 2 is set. Let's get together. Similarly, the seventh horizontal evaluation value generating circuit 6 evaluation value E ₇ generated by 2 g of the value of c fin dough W 1 is set, the object

Since the high frequency component of B is included, it is not appropriate as the evaluation value of the object Λ. Therefore, the evaluation value E ₇ is rather considerably larger than the value of the evaluation value E ₈ the value of the windowing W 2 is generated by the eighth horizontal evaluation value generating circuit 6 2 h of which is set Natsute I will.

Also, some window fin dough W 2 in INO such exist NonTarget object B, and the evaluation value E ₂ or the evaluation value E ₈ is say whether it is appropriate, it can be said that no means appropriate. The reason is explained with reference to Fig. 18B. FIG. 18B shows a case where the lens is moved so that the force is adjusted to the object Α. The more the focus is adjusted for the vacation A, the more the focus for the object B is shifted. If the focus of the object B is largely displaced, the image of the object B is largely blurred, and the blurred image enters the window W2. Therefore, in the case of the states of FIGS. 8A and 18B, the evaluation TE ₂ generated by the second horizontal evaluation value generation circuit 62b in which the value of the window W2 is set is set. ¾ is by no means appropriate. Similarly, the value of the eighth horizontal evaluation t generator 6 2 evaluation value generated by h E ₈ the value of the windowing W 2 is set can not be said to be in any way appropriate. Thus, for windowing W 1 by the obtained evaluation value E, and E _7, window W 2 Niyotsu evaluation obtained Te E ₂ and E 8 determines whether it is appropriate, the

I,,-Ε 2 I ≤ E, x; 9

And ♦ · · · (3)

I Ε 7-Ε ₈ I ≤ Ε 7/5

Should be determined. Note that a coefficient set in advance by experimental results, here, for example, β = 0. 0 1 but there is a, (E, X beta) and (E ₇ X beta) and (3) Without using it in the formula,

(E, X beta) and (E ₇ X beta) the results using a predetermined value experimentally obtained in (3) in place of the same.

The (3) of the result in based determination, | Ε, - Ε ₂ I and I E 7 - if the value of E 8 I is less than the predetermined value, the evaluation value E, the difference between E ₂ question is it can be determined that little, and the difference in ίΙΟ evaluation value E ₇ and E ₈ it can be determined that little. Therefore, it is determined that there is no non-target rest B as shown in Fig. 18Λ. If the values of IE,--E ₂ I and IE ₇ — E ₈ i are equal to or greater than a predetermined value, it can be determined that the difference between the evaluation values E, and E 2 is large, and the evaluation ¾Ε ₇ and Ε _δ It can be determined that the difference is large. Therefore, it is determined that there is a non-target object as shown in Fig. 8 {Fig. That is, (3) calculates the equation, if the fit this equation (3), the evaluation value E, and E 2, the evaluation value E ₇ and E ₈ are proper, against this equation (3) or If so, the evaluation values 評価 and Ε ₂ and the evaluation values F and Ε ₈ are inappropriate.

Next, steps S105 to S5 will be specifically described with reference to FIGS. 10 and 11 in consideration of the above-described basic idea.

In step S105, the lens position is X.時に obtained at the time of

3 8 (X 0)-Using E ₂₄ (X.)

I E, (X o)-E 2 (X o) I ≤ E, (X.) x, and

IE ₇ (X.) 1 E ₈ (X o) ≤ E (X o) x β 1

(105) is determined. Evaluation value E,, E _2, E ₇ and E ₈ is, if the (1 0 5) than a value that applies to expression evaluation value E,, E _2, E 7 and E ₈ are the appropriate values It proceeds to step S117. Conversely, if the evaluation values Ε, Ε 2, Ε 7, and で₈ are values that do not apply to this (1 0 5) equation, at least the evaluation values Ε,, Ε ₂ , Ε ₇ and E ₈ proceeds to stearyl-up S 1 eta 6 it is determined that the inappropriate value.

In Step S 1 0 6, thus evaluation value calculation result in step S 1 0 5 E, and E _2, the evaluation value E ₇ and E 8 is determined to be inappropriate, following windowing W 1 The evaluation values Ε ₃ and Ε ₃ obtained based on the larger window W 3 are used, and the evaluation values Ε ₄ and ₄ obtained based on the next larger window W after the window W 2 are used. Ε,. Use

In step S106, the lens position is X as in step S105. Using Ε, (X 0) E ₂₄ (X.) obtained at the time of

IE ₃ (Χ.) I E ₄ (X o) I ≤ E (X o) x β and

IE ₉ (X.) One E,. (X.) ≤ En (X o) x β _t

(106) is determined. Evaluation value E _3, E 4. E _s and E,. But if it's the value that applies to the (1 0 G) expression evaluation value E _3, E 4, E n and E,. Is determined to be an appropriate value, and the process proceeds to step S107. Conversely, the evaluation value E _3, E _4, E ₃ and E _{l 0} is, if the (1 0 6) of a applies should not be a value in the expression, at a minimum, the evaluation value E _3, E _4, E ₃ and E,. Is determined to be an inappropriate value, and the flow advances to step S108.

Here, the reason why the larger windows W3 and W4 are used will be described. As described above, in the state indicated by the 1 8 lambda diagram, the evaluation value E, and E _2, the evaluation value E ₇ and E ₈ are improper, the target object A or the non-target and off O carcass It cannot match any of the objects B. However, if the larger windows W3 and W4 are used for the windows W1 and W2, the non-target rest B may enter the window W4. You could think so. If completely when the non-target object B ever fall within the © Lee down dough W 4, the difference is with small Kunar between evaluation values E ₄ and the evaluation value E _3, the evaluation value E ₃ E,. The difference between is smaller. In other words, the evaluation values E ₃ , E ₄ , E ₃ and E, ₀

) Will be the value that applies to the expression. As a result, since the evaluation values E ₃ , E 4, E 9, and E, ₀ are appropriate values, the focus matches the non-target object B. Originally, the focus should be adjusted to the target object A, but if you try to adjust the focus to a vacation, an appropriate evaluation value cannot be obtained.

Then, the auto focus control circuit 34 repeats the control loop many times, and keeps moving the focus lens for a long time. Therefore, the problem that the loop is repeated and the blurred imaging signal must be continuously output. However, by focusing on the non-target object rest B, the control loop is repeated for a long time to prevent the output of the blurred imaging signal from being continued. Can be.

In step S107, the evaluation value determined in step S10 E,, E ₂ , E 7, and E ₈ are inappropriate values, and the evaluation values E ₃ , E 4, E a, and E ₁₀ determined in step S 106 are appropriate values. Based on the result, there are defined only the numbers i = 1, 2, 7, and 8 as unused numbers, and the process proceeds to step S117. I = 1 the unused number in Step S 1 0 7 This, 2, 7, because it was 8 Gajo defined, at step S 1 0 7 subsequent step, the evaluation value E,, E _2, E ₇ and it is not the E ₈ is used.

In step S 108, the evaluation values E ₃ and E ₄ , the evaluation values E ₃ and E, based on the calculation result in step S 106. Was determined to be inappropriate, the evaluation values E ₅ and E,, obtained based on the next largest window W 5 after the window W 3 were used, and the window W 4 then use a large Huy down de W evaluation value E ₆ was obtained based on the 6 and E, _2.

In step S108, the lens position is X as in step S106. Using E (X E 24 (X o) obtained at the time of

IE ₅ (X.) One E ₆ (X o) E r, (X 0) X ^, and

I E (X E, 2 (Xo) E,, (X „) X,

(10 8) is determined. The evaluation values EEE, and E, ₂ are

) If the values satisfy the expression, it is determined that the evaluation values E, f;, EI 1 and E, ₂ are appropriate values, and the flow advances to step S 109. Conversely, if the evaluation values E ₅ , E _G , E,, and E ₁₂ are values that do not apply to the expression (108), at least the evaluation values E ₅ , E 6, E,, And E and ₂ are determined to be inappropriate values, and the process proceeds to step S110. In Step S 1 0 9, the evaluation value E, it is determined at step S 1 0 5, the result of E _2, E 7 and E ₈ are improper values, it is determined by the scan Tetsupu S 1 0 6 evaluation value E _3, E 4, E 9 and E ₁₀ is a result of an unsuitable value, step S 1 0 8 evaluation value is determined by E _5, E 6, E and E ₁₂ is a proper value Based on the result, only the numbers i = l, 2, 3, 4, 7, 8, 9, and 10 are defined as nonuse signs, and the process proceeds to step S117. In this step S109, i = 1, 2, 3, 4, 7, 8, 9, 9 and 10 were defined as non-use cases, so in the steps after this step S109, evaluation value _{_{E,, E 2, E 3}} , E 4, E 7, E 8, E 9及beauty E _{l 0} is never used.

In Step S 1 0 8, Step S 1 0 Thus evaluation value calculation results at 6 E ₃ and E _4, the evaluation value E: ₃ and E, because η is determined to be inappropriate, Huy down dough W 3 evaluation value obtained based on the large Ui down dough W 5 next £ ₅ and with using the E, evaluation value obtained based on a large windowing W 6 in the following Ui down de W 4 E ₆ and to use the Ε _12.

In step S11, the lens position is X, as in step S11. Using Ε, (Χ.) ~ Ε ₂₄ (Χ _η ) obtained at

| Ε _{Ι 3} (Χ.) Ε, 4 (Χ.) Ε (X.) x β 2 and

1 Ε ₁₃ (Χ.) One E _2. (Χ.) Ε (X 0) X β 2

(1 1 0) is determined. If the evaluation values Ε, ₃ , Ε ₁₄ , Ε _13, and Ε ₂₀ are values that apply to this (1 110) equation, the evaluation values Ε ₁₃ , Ε, ₄ ,, 9, and Ε ₂₀ are appropriate. The value is determined to be a value, and the process proceeds to step S111. Conversely, the evaluation value _{_{E l 3, E l 4,}} E l 3 , and E ₂₀ is, in this (1 1 0) Formula If it's a true no value, at a minimum, the evaluation value _{_{E l 3, E l 4,}} E, 3 , and E _2. Judge to be an inappropriate value and proceed to step S112? _ ₀

In step S 1 1 1, evaluation values E, it is determined at step S 1 0 5, the result of E _2, E 7 and E ₈ are improper values, it is determined by the scan Tetsupu S 1 0 6 Evaluation values E ₃ , E ₄ , E 9 and E,. Is an inappropriate value, and the evaluation values E ₅ , E ₆ , E,, and E ₁₂ determined in step S 108 are inappropriate values, and 1 evaluation value is determined in the _{_{0 E l 3, E, 4}} , E 13 and E ₂₀ is based on the result that a suitable value, only the i =. 1 to 1 2 number as a non-use number Define and go to step S 1 17. Since i ==. 1 to 1 2 is defined as non-use numbers in this step S 1 1 1, in step S 1 1 1 subsequent step, the evaluation value E, ~ Ej ₂ is to be used There is no.

In step S112, the lens position is X as in step S110. Using E, (X.) to (X.) obtained at the time of

I E, 5 (Χ θ)-E, 6 (X o) E, 5 (X o) x β

And

1 iiv 2 1 (X 0)-li 22 (X 0) I ≤ E ₂₁ (X o) x ^ 2

Judge (1 1 2). Evaluation value _{_{E, 5, E 16, E}} 21 and E ₂₂ are, if the (1 1 2) than the value applicable to the expression evaluation value E,, E _{l 6,} E 2, and E ₂₂ are proper _π Conversely proceeds to step S 1 1 3 it is determined that the value, the evaluation value E _15, E, _6, E _{2 iota} and E ₂₂ are, in this (1 1 2) applies should not be in the expression i if located in, at a minimum, the evaluation value E, ^ E _IG, the process proceeds to step S 1 1 determines that the E ₂₁ and E ₂₂ are improper values. In step S 1 1 3, evaluation value E, it is determined at step S 1 0 5, the result of E _2, E 7 and E ₈ are improper values, it is determined by the scan Tetsupu S 1 0 6 The evaluation values E ₃ , E 4, E 9, and E _η are inappropriate values, and the evaluation values E _s , E 6, E,, and E ₁₂ determined in step S ₁₀₈ are incorrect. a result that is an appropriate value, a result that is inappropriate value is determined evaluation value _{_{E l 3, E l 4,}} E, 3 , and E ₂₀ at step S 1 1 0, step S 1 1 2 in钊断evaluation value _{_{E l 5, E, 6,}} E 21 and E ₂₂ based on the result that it is an appropriate value, i = l~ 1 4, 1 9, 2 0 of respondent No. only Define as an unused number and go to step SI17. Since i = 1 to L2, 19, and 20 are defined as unused numbers in step S113, the evaluation values E, to E, _4, E _13, E ₂₀ is never used.

In step S114, the lens position is X as in step S110. Using E, (X.) to E ₂₄ (X.) obtained at the time of

| E, ₇ (X o) -E ₁₈ (Xo) E, 7 (X o) β 2 and

| Ε ₂₃ (Χ.) Ε ₂₄ (Χ.) Ε ₂₃ (X ο) X β 2

Judge (1 1 4). The evaluation values Ε ₁₇ , _{Ι Ι 8} , Ε ₂₃ and Ε 2 are (1 1 4

) If the a value applicable to the expression evaluation value _{Ε, 7, Ε 1Β, Ε} 23 and E ₂₄ are determined to be a suitable value the process proceeds to step S 1〗 5. Conversely, if the evaluation values E | ₇ , E _l8 , E _23, and E ₂₄ are values that do not apply to the equation (114), at least the evaluation values E _{l 7} , E _{l 8,} E ₂₃ and E ₂₄ are determined to be an unsuitable value proceeds to stearyl-up S 1 1 6.

In step S115, the evaluation value determined in step S105 The results that E,, E ₂ , E 7, and E ₈ are inappropriate values, and the evaluation values E ₃ , E 4, E 9, and E, determined in step S 106. And results because the result of an inappropriate value, the step S 1 0 evaluation value is determined in the 8 E _5, E 6, E and E ₁₂ are improper values, determined in step S 1 1 0 evaluation value _{_{E l3, E l 4, E}} , 3 , and

E ₂₀ is a result of an unsuitable value, a result that would have Step S 1 1 2 in the determined evaluation value _{_{E, 5, E 16, E}} 21 and E ₂₂ is an inappropriate value, stearyl Tsu Based on the results that the evaluation values E _{l 7} , E, ₈ , E ₂₃ and E ₂₄ determined in step S 1 ₁₄ are appropriate values, i = 1 to 16, ₁₉ to 22 Define only the number as an unused ban, and go to step S117. Since i = 1 to 16 and 19 to 22 are defined as non-use numbers in step S115, the evaluation values E, to E in step S115 and subsequent steps are defined. , _G, never E _{I 3} to E ₂₂ is used.

Reaching step S 1 16 means that all evaluation values E, to E

_{Since 24} is inappropriate, auto focus operation is determined to be impossible, and the mode shifts to manual focus mode, and the processing ends.

This concludes the steps for selecting an inappropriate evaluation value from the 24 evaluation values.

Steps S〗 17 to S1331 shown in FIGS. 12 and 13 are specific operation steps π- for determining the lens moving direction. Steps SI 17 to S 13] are steps performed by the micro-computer 64.

In step S117, i is set to 1 and the up force value Ucnt, the downcount value Dcnt, and the flat count value Fcnt are reset.

In step S118, i is defined as an unused number It is determined whether it is a number. If i is not defined as a non-use waiver, go to step S120. If i is a number defined as an unused ban, i is incremented in step S119 to proceed to the next i number. In Step S 1 2 0, E i ( X 0) is, rather than the value of the same extent as E i (X _2), has a value greater than a certain degree E i (X _2), one且, E This is a judgment when i (X,) has a value that is not the same as Ei (X.) but is somewhat larger than Ei (Xo). To explain more clearly, when the focus lens moves in the Far direction such as X ₂ , X 0, X, the evaluation value becomes E, (X ₂ ), E i (X ₀ ), E i (X,). In particular,

Ei (Χ ₂ ) x β Ε Ε i (X ο)

And

Judgment is made by performing the operation of (120). Here, β ₃ is a coefficient obtained experimentally, and is set to / 5 ₃ = 1.0 3 in the present embodiment. The fact that the condition of the expression (120) is met means that the focal lens is Χ ₂ , Χ. , X, and the like, means that the valuation increases in order, and the process proceeds to the next step S122. If the condition of the equation (120) is not satisfied, the weight data Wi is added to the upcount value Ucnt in step S122 in step S122, and the value in step S126 is calculated. move on.

In step S 122, E i (X 0) is not the same value as E i (X,), but has a value that is somewhat larger than E i (X,). i (X ₂ ) is not the same value as E i (X o), This is the case when the value has a value larger than the degree E i (X.). Further understanding easily explained, off O carcass lenses, X _2, X 0

, X, when moved in the F ar direction so on, evaluation value _{E i (X 2), E} i (Χ 0), it is determined that whether or not down to the order of E i (X,). In particular,

E i (X,) x β <E i (Xo)

And

E i (Xo) X β <E i (X ₂ )

It is determined by performing the operation of (1 2 2). The fact that the condition of equation (120) is met means that the focus lens is X ₂ , X. , X, which means that the evaluation values are sequentially reduced, and the process proceeds to the next step S123. If the condition of the equation (120) is not satisfied, the process proceeds to step S124.

In step S123, the weight data Wi is added to the downcount value Dcnt, and the process proceeds to step S126.

In step S 124, E i (X.) does not have the same value as E i (X,), but has a depression that is somewhat larger than E i (X,). i (chi ₀₎ is, rather than the value of the same extent as E i (X 2), which is determined when it has a value greater than a certain degree E i () _2). Further understanding easily explained, off O over mosquito scan lens, X _2, X 0

, X, and so on, the peak of the evaluation value is at E i (X o). As a rest,

E i (X,) X β 3 <E i (X o)

And

E i (X ₂ ) x β ₃ <E i (X o)

It is determined by performing the operation of (1 2 4). When the condition of equation (1 2 4) is satisfied, That matches the focus lens, X _2, X. , X, means that the peak of the evaluation value is at E i (X o), and the flow advances to the next step S 125. When the condition of the equation (120) is not satisfied, the process proceeds to step S126.

In step S125, the weight data Wi is added to the flat count value Fcnt, and the process proceeds to step S26.

In step S 1 26, i is incremented and step S 1

Proceed to 2 7.

In step S127, since two evaluation values E are generated in the horizontal evaluation value generation circuit 62 and the vertical evaluation value generation circuit 63, it is determined whether i is 24. If the power is 24, it is determined that the calculation for all the evaluations has been completed, and the process proceeds to step S128. If i is not 24, the loop composed of steps S118 to S127 is repeated until i becomes 24.

In step S128, the upcount value Ucn, the downforce value Dcnt, and the flatcount value Fcnt are compared to determine which It is determined whether the default value is the largest value. If the upcount value Ucnt is the largest value, the process proceeds to step S129, and if the downcount value Dent is the largest value, the process proceeds to step S129.

Proceed to step S30, and if the flat count value Fcnt is the largest value, proceed to step S131.

In step S129, the micro computer 64 determines that the direction of X, is the hill-climbing direction of the evaluation value, that is, the direction in which the focus is in conformity. And then follow the direction of Far. .

In Step S 1 3 0, microphones b Computer 6 4, chi ₂ hill-climbing direction of the direction evaluation value, ie, in the direction of full Okasu fit Judgment is made, and the CPU 4 specifies the lens movement direction as the Near direction.

At step S131, the microcomputer 64 is X. The position is determined to be the position where the focus is adjusted, and the process proceeds to step S2188.

The operation of the above-described steps S118 to S131 will be described in an easy-to-understand manner with reference to the example shown in FIG. The first 9 figures lens position X _2, X. FIG. 7 is a diagram showing, as an example, a transition of changes in evaluation values E i (X ₂ ), E i (X ₀ ), and E i (X,) in (,, ().

First, in step S118, it is determined whether or not i is an unused Φ number. Here, it is assumed that all i can be used.

First, in the first loop, a decision is made to check for E,

E, (X ₂₎ Ku E, (X o) <E , because it is (X,), consistent with the scan Tetsupu S 1 0 of the conditions, the process proceeds to step S 1 2 i.

Therefore, in step S122, the operation of Ucnt-0-W, is performed.

In the second loop, a determination is made that閿to E _2,

Since E 2 (X ₂ ) and E 2 (X o) <E ₂ (X,), the condition of step S 120 is met, and the process proceeds to step S 12〗.

Accordingly, Ucnt in Step S 1 2 1 - W, the operation of the -I- W ₂ is carried out.

In the next third, fourth, and fifth loops, the same operations as in the first and second loops described above are performed, and in step S12 of the 50th loop,

Ucnt two _{W, W 2 + Wn 4 W} , + W Γ, operation is performed.

In the sixth loop, a determination is made on Ε _2, Since E 2 (X ₂ ) <E 2 (X o)> E 2 (X,), the condition of step S 124 is met, and the flow proceeds to step S 125.

Accordingly, in step S125, the calculation of Fcnt = 0 + Ws is performed.

After repeating the loop 24 times in this way, eventually,

U cnt-W, + W ₂ + w ₃ + W ₄ + W ₅ + W ₇ +

W ₈ + W ₉ + w,, + W, 3 + W, ₄ + W. 5 +

F cnt = w ₆ + W. 2 + W, 9

The calculation indicated by is performed. Therefore, when the values of the weight data W i shown as an example in FIG. 7 are substituted for the up-count value Ucnt, the down-count value Dent, and the flat-count value Fent,

U cnt = 1 2 4

D cnt = 1 3

F cnt = 1 8

Is the result.

Accordingly, in the determination in step S128, the upcount value Ucnt has the largest value, and in the example shown in FIG. 9, the process proceeds to step S129. Therefore, the force direction is determined to be the X, direction.

The following steps S200 to S222 are steps for determining the lens position at which the evaluation value is maximized. This operation flow is an operation flow executed by the microphone computer 64. Steps S200 to S221 are performed, and FIG. 3 to FIG. 15 are referred to. I will explain in detail.

Here, the following step S200 and the following explanations will be clarified. for,

X! X o + Δ X

X 2 X 0 + 2 X Δ X

X 3 X 0 + 3 X Δ X

X κ = X 0 + k X A X

X _{k + I} -X o + (k + 1) x Δ X

X _{k +} j = X ο + (k -fj) x Δ X

(2 0 0) and: {1}.

In this embodiment, since the evaluation value is sampled in one field ¾, the distance indicated by ΔX is the distance that the focal lens moves between one field. Is defined. Therefore, δΐί separation Δ を represents the distance the lens travels in one field period. The distance △ X not only represents the distance the lens travels in one field period no, but also is based on the lens movement direction obtained from step S100 to step S13ϋ. Thus, the polarity of ΔX is determined. For example, if the lens movement direction is the Far direction, Δ Δ is set to have a positive polarity, and if the lens movement direction is the Near direction, △ X is the negative polarity. Is set to have

First, in step S200, k = 1 is set.

In step S 2 0 1, Mai Croc Npyu Ichita G 4 are instructs to move the lens to the position of the X _k to the CPU. Here, the lens position X _k is given by the following equation (2 0 0).

X _k = X 0 + k X Δ X

I will go to the beach.

In step S202, the microcomputer 64 receives the evaluation values E, (Xκ) to be newly generated in the horizontal evaluation value generation circuit 62 and the vertical evaluation value generation circuit 63. The evaluation value E ₂₄ (Xκ) is stored in the RAM 66. The 24 evaluation values E i are stored in a table as shown in FIG.

In step S203, set i = 1, j = 1, and

Five

Reset the up count value U cn t, down count value D cn t., And flat force point value F cnt.

In step S204, it is determined whether or not i is a ban which is defined as an unused number. If i is not defined as an unused number, go to step S206. If i is a number defined as an unused number, step S 2

In 0 5, i is incremented, and the people are returned to step S 204 again.

In stearyl-up S 2 0 6, off O Kasurenzu is, X _k, evaluation value ^ was when moved to the position indicated by X _k from E i (X) is, the evaluation value E i (X _k _, ) Is judged to be higher than a certain level. In particular,

E i (X _k _,) x β ₄ <E i (X k)

It is determined by performing the operation of (206). Here, β is a coefficient obtained experimentally, and in this embodiment, ₄ = 1.05. The fact that the condition of the equation ( ₂₀₆ ) is met indicates that the evaluation value E i (X _k ) is more than the evaluation value E i (X,) to some extent. In this case, the next step S 2 0 7 Proceed to. When the condition of the equation (206) is not satisfied, the process proceeds to step S209.

In step S207, the evaluation value E i (X _k ) force ^ Since the evaluation value E i (X _k —,) has been increased to some extent or more, the UZ D information (up Z down information) is used. Then, the 2-bit data “01” indicating that the data is up is stored in the RAM 66 in association with the evaluation value E i (X,).

In step S 208, similarly to step S 122, weight data W i is added to up force event value U cnt, and the flow advances to step S 214.

In step S 2 ϋ9, the evaluation value E i (X _k ) obtained when the focus lens moves from X _k —, to the position indicated by X becomes the evaluation value E i (Χ,- A judgment is made as to whether or not,) is down to some extent. In particular,

E i (X _k ) x β 4 E E ₄ (X,-,)

It is determined by performing the operation of (209). The fact that the condition of the expression (209) is met indicates that the evaluation value E i (X _k ) is down to a certain extent with respect to the evaluation (X _k −,). In this case, the process proceeds to the next step S210. When the condition of the equation (209) is not satisfied, the process proceeds to step S212.

In step S210, since the evaluation value E i (X _k ) is down to a certain extent with respect to the evaluation value E i (X _k- >), the IJZ I) information (up / down information) Then, 2-bit data “10” indicating down is stored in RΛ66 together with the evaluation value E i (X _k ).

In step S211, as in step S123, the down-force value Dcnt! ! Step S2 1 4 Proceed to.

The fact that the step S212 is reached means that the focus lens is represented by X _k —, to X _k , considering the conditions of the step S 206 and the conditions of the step S 209. It means that the evaluation value E i (X _k ) obtained when moving to the position where the evaluation value E i (X _k _,) has not changed by more than a certain degree .

Therefore, in step S212, 2-bit data “00” indicating a flat is used as UZD information (up-Z down information) as an evaluation value E i (X _k ). Relate and store it in R 6 66.

At step S213, similarly to step S125, the weight data W i is added to the flat count value Fcnt, and the process proceeds to step S2J4.

In step S2114, i is incremented, and the process proceeds to step S215.

In step S215, it is determined whether or not the i force is 24. If i is 24, it is determined that the calculations for all the evaluation values have been completed, and the flow proceeds to step S216. When i is not 2/1, the loop composed of steps S204 to S21Γ) is repeated until i becomes 24.

Step S216 is a step for determining whether or not the downcount value Dcnt has the largest value.

Here, taking FIG. 20 as an example, this step S 閲 6 will be described. FIG. 20 is a diagram showing a storage state of each evaluation data and each up-Z down information at RΛΜ66. As shown in FIG. 20, the micro computer (; 4) associates each evaluation i with each up / down information so as to correspond to the moved lens position κ κ. 6 to memorize. First, when the lens position is chi _kappa, when step S 2 0 4 crows Tetsupu S 2 1 5 loops are repeated, Appuka window down preparative value U cnt, Dow linker window down preparative value Dent and off rats The count value F cnt is as follows.

U cnt = W, + W ₂ + w,, -f W ₅ + W, ₃ + W + w,! 4 W. 4 + w, 15 + w, 6 + w, I 9 + W 2 3

D cnt = W ₇ + W, 0 + w, I 7 + w, 8 + W:> 0 + w 2 I-

F cnt = W ₃ w ₆ + Vv ',, 2 + w, 3-W ₂ Ϊ 2

For this up-count value U cnt, down-force value D cnt and flat-count value F ciU, the value of the weight data Wi shown as an example in に 7 is Substitution gives

Ucnt = 9 5

D cnt = 3 4

F cnt = 3 1

Is the result. In other words, from these results, it can be said that the individual evaluations ft: vary in ups, downs, flats, etc., but the evaluation values increase if they are comprehensively compatible. You can judge.

Hereinafter, the evaluation value comprehensively determined in step S216 in this manner will be referred to as an “overall evaluation value”. Therefore, if step S216 is expressed in another word, step S216 is a step for judging whether or not comprehensive evaluation 值 is down. Can be expressed as

Next, consider the case where the lens position is Χ «, as shown in FIG. When the lens position is X _{K tl,} the Le one flop of Step S 2 1 5 from stearyl-up S 2 0 4 is repeated, Ryo Ppuryokuu down preparative value U cnt, Dow Nkau down preparative value D cn t And the flat count value Fcnt are as follows. Ucnt = W ₅ + W,, + W, 2 + W, 7 + W, ₈ + W ₂₀ + W ₂₃

Dent = W, + W ₂ + W ₃ + W ₆ + W ₇ + W ₈ +

W, 0 + W, 3 + W, 5 + W ₁₆ -W _{l 9} +

F cnt = W ₄ -W o

Substituting the value of the weight data Wi shown as an example in FIG. 7 into the upcount value Ucnt, the downcount value Dr, nt, and the flatcount value Fcnt. ,

Five

U cnt-2 9 7

Dent = 1 1 3

F cnt = 1 8

Is the result. In other words, from this result, it can be determined that the overall evaluation value is down. Thus, when it is determined in step S216 that the overall evaluation value is down, the process proceeds to step S217.

In step S217, j is incremented, and the process proceeds to step S218. This j is a value indicating how many times the judgment result in step S216 has become YES, that is, a value indicating how many times the total evaluation has been reduced.

In step S 218, let を_{κ + を} be the first lens position of the lens position at which the comprehensive evaluation value starts to be continuously reduced. In this step S ₂₁₈ , X κ Judge whether the moving distance of the lens from (Xκ _{+ j} is greater than DXn. The formula for making the actual judgment is

Δ XX j ≥ DX n (2 18). Note that D represents the focal depth of the focus lens, and n represents a preset coefficient. However, according to the experimental results, When set in the range of 1 ≤ n ≤ 10, autofocus operation can be realized at an optimum speed.

Here, the judgment operation of step S218 will be described with reference to FIG. The axis of 橫 in FIG. 21 represents the position X of the lens, and the axis of ordinate represents the evaluation value E (X) for the position of the lens.

When j = 1, the lens position corresponding to j = 1 is X _{κ +} , since the overall evaluation value is the lens position when the first down was performed. Therefore, the right side (△ XX j) of the equation (2 18) is the lens position 総合_{κ + at which the} overall evaluation value first starts to decrease from the lens position XK before the overall evaluation value starts to decrease. Represents the distance between and. However, as can be seen from FIG. 21, the judgment result in step S218 is 8◦.

When j = 2 is because it is the lens position when the comprehensive evaluation value goes down with two consecutive lens position corresponding to j = 2 is X _2. Therefore, from Fig. 21, the right side (Δ} (X j) of the equation (2 18) is calculated from the lens position X κ before the comprehensive evaluation value starts to go down.

Represents twice lenses position distance X _{K 4 2} began to down continuously. However, as can be seen from FIG. 21, the judgment result of step S218 is Ν0.

In the case of j-3, as in the case of .i2, the judgment result in step S218 is NO.

When j = 4, the lens position corresponding to j = 4 is X because the comprehensive evaluation value is the lens position of した that has been down four times in a row. Therefore, from Fig. 21, the right side (辺 XX j) of the equation (2 18) is the lens position before the total evaluation value starts to decrease. It represents the distance of the lens position X _{K f 4 where} the lens started to move. Therefore, as can be seen from FIG. 21, (△ XX j) ≥ Γ) X n, and the result of the determination in step S 218 is YES.

On the other hand, in step S216, the downcount value Dcnt is If it is determined that the total evaluation value does not have the largest value, it is determined that the overall evaluation value has not decreased, and the process proceeds to step S219.

In step S219, j = 0 is set. This is the step to reset j. The reason why j is reset is that j is a value indicating how many times the overall evaluation value has been reduced. More specifically, reaching step S219 means that the overall evaluation value was determined to be not down in the determination result of step S216 ' Therefore, in the determination in step S216, the continuous down of the total evaluation value is stopped. Therefore, j is reset in this step S219.

When the continuous down of the total evaluation value is interrupted, j is reset. For example, in Fig. 2, a certain evaluation value E (Xκ) is a local maximum value due to mere noise. Even if this is the case, j is reset in the operation loop of the evaluation value E (Χκ ₊ ,) or Ε (XΚ.2) or Ε (Xκ, 3), so that the evaluation value E (X κ) is not recognized as the maximum.

In step S220, k is incremented in order to further move the focus lens, and the process returns to step S221.

When the determination result of step S2 18 becomes Y E S, step S2

2 Proceed to 1. In step S22], it means that the total evaluation has been continuously downed from the lens position Xκ for a predetermined number of times (j times). Therefore, the micro computer 64 receives the lens position Xκ. Is determined to be the lens position X _{g at} which the evaluation value is maximized.

Next, from the evaluation values E i (X κ) at the lens position Χ _κ ,

Based on the up-down information of R Λ M66, the overall evaluation value and the up-down state near the lens position X κ and the up-down of Ei stored in the RAM 66 Select the i that matches the Just do it. Assuming that the selected weight i has the highest weight data Wi is W _g , the maximum evaluation value is defined as E _g (Χ κ)

. When the maximum evaluation value E g (Χ _Κ ) is defined, the corresponding lower limit evaluation value is defined as E _g (Χ κ ₊ ,). The maximum evaluation value Ε ₉ (Χ _κ ) is updated every field even after focusing has been achieved by fixing the lens to Χ _κ wide, but the lower limit evaluation value is E _g (X κ ₊ .) is fixed.

This is explained in the example shown in FIG. When the lens position is _{κ κ} , the overall evaluation value is increased based on the judgment of step S 2 16, and when the lens position is X κ ₊ , step S 2] 6 Based on this judgment, the overall evaluation value is down. Therefore, huh. In the example shown in Fig. 20, the down information is up when the lens position is XΚ and down when the lens position is X _{κ +} , i = l, 2, 5, 8, 14, 15, and 19. Among them, the weight data W i is highest according to the data in FIG.

Since it is 1, E, (Χ _κ) is used as the maximum evaluation value. In stearyl-up S 2 2 2, microphones Computer 6 4 evaluation value sends a control signal to the CPU 4 so as to move the full O carcass lens to the maximum and a lens position X _g.

In step S223, it is determined whether or not the CPU 4 has received a command to follow the subject. The subject tracking command is used to control the tilt and pan movements of the video camera so as to follow the movement of the subject, and to adjust the position of the evaluation value detection window of the video camera's unique focus. This is a command to make it variable. For example, the following command is supplied to the CPU 4 when the camera man presses a following command button provided on the operation unit 5. If the following command has been supplied from the operation unit 5, the process proceeds to step S300. If there is no following command, the process proceeds to step S224. In step S224, it is determined whether or not an autofocus stop command has been issued. If the autofocus mode has been released by the operation of the camera, the flow advances to step S225 to shift to the manual focus mode.

If there is no instruction to stop the autofocus mode in step S224, the process proceeds to step S226, where the maximum evaluation value E _g (Xκ) and the lower limit evaluation A comparison of the values E _g (Χ κ ₊ ,) is made. If the value of the maximum evaluation value E _g (X κ) decreases due to a change in the subject and becomes lower than the lower limit evaluation value Ε ₉ (Χ _{κ + 1} ), the process proceeds to step S 2 27. Restart the to focus operation. Talent

—If the scrap operation is restarted, return to step S100 again. Ο

Next, the recognition operation of the set subject will be described with reference to the flow from step S3H0 shown in FIG. The flow from step S300 shows the processing operation performed by the CPU 4. Further, in order to more specifically I3 ¹ solution the description of this flow, at the same time, also the reference example shown in the second 2 FIG.

FIG. 22 shows a state in which a round object Λ and a square object Β are being imaged. In this example, it is assumed that the color of the object と and the color of the object rest match the color set as the target object. still

As shown in Fig. 22, the origin is the raster scan start point (upper left in the imaging screen), and the horizontal scanning direction is defined as the X-axis direction, and the vertical scanning direction is defined as the Y-axis direction. Therefore, on the imaging screen, the coordinates of the raster scan start point are (0, 0), and the coordinates of the raster scan end point are (768, 240). The center of (3

84, 120).

In step S223, the CPU 4 receives the motion tracking instruction of the subject from the operation unit 5, and receives the command from the operation unit 5 in step S3 shown in FIG. 0 Move to 0.

In step S300, the camera man operates the operation unit 5 to determine whether or not a subject as a target object has been set. Here, a method of setting a subject as a target object will be described. Cameraman first captures a desired subject as a target object so that it is located at the center of the imaging screen. Next, with the target object positioned at the center of the imaging screen, when the camera man presses the target object rest determination button provided on the operation unit 5, the CPU 4 causes the object located at the center of the imaging screen to move. Is recognized as the desired target object of the force camera, and the color information of this object is loaded into the RAM 4a. The setting of the target object is not limited to this method

Alternatively, an object having a preset color (for example, flesh color) may be set as a target object. In step S3ϋ0, when the color information of the subject set as the target object is stored in RΛ4a, the flow shifts to step S301.

In step S301, CP IJ4 determines the object set in step S300 from the four object modes (modes 0, 1, 2, and 3) described in. Select the most suitable subject mode for the object. The CPU 4 controls the switching of the switch circuits 72a.72b.72c and 72d according to the selected subject mode. In this step S301, when the object mode is selected by the CPU 4 and the switching of the switch circuits 72a, 72b, 72c and 72d is controlled, The search circuit 38 searches for an area in which rain element data that matches the color component of the subject set as the target object exists. This search processing is not the processing performed by the CPU 4, but an area search circuit provided as a hardware circuit.

It is an essential point of the present invention that the process 38 is performed. In other words, by configuring the area detection circuit 38 as a hardware circuit in this way, all pixel data from the encoder 37 can be processed. This means that condition judgment processing can be performed in real time.動作 The operation of the relay search circuit 38 has already been described with reference to FIG.

In step S302, the CPU 4 determines which area has the pixel data of the same color as the target object based on the chip circuit number supplied from the end search circuit 38. recognize. Thus, the CPU 4 can select only an area having the same color as the subject set as the target object. In the example shown in the second 2 figure as a d re A present the same color as the object object, lambda _{0 6 8}ゝA 0 6 9, A 0 8 4, A 0 8 5ヽA 0 8 6 , A 0 β 7 χ Λ 10 2 and A 10 3 will be selected.

In step S303, the CPU 4 reads out all pixel data present in the area selected in step S302 from the frame memory 39 in the last scan order. The pixel data existing in the area not selected in step S302 is not read at all. By supplying the same address from the frame memory 39, the pixel data is read out into the pixel data composed of Υ data, R-Υ data and Β-Υ data. In the example shown in FIG. 2, CPU 4 has eight areas A 0 _{fi 8} , A 0 6 9 A 0 8 4 A 0 8 5 A 0 8 6 ゝ Λ 0 B 7 Λ Λ] 0 2 and Λ ΐ 0 3 ϊ Only the existing pixel data is read from the frame memory 39. At this time, the images existing in the areas other than the selected eight areas, f, are not read out from the frame memory 39 at all. As described above, the CPU 4 determines the area from which the pixel data is read from the frame memory 39 based on the result searched by the area search circuit 38. The number of pixel data that! Can receive from the frame memory 39 can be reduced. Therefore, CPU 4 is in real time, and the entire image supplied from frame memory 39 is Processing can be performed on raw data.

In step S304, the CPU 4 first selects the mode 0 as the subject mode based on the read pixel data including the Y data, the R-Y data, and the B-Y data. If so, the condition is determined based on the equation (700). Mode as subject mode

When 1 is selected, a condition is determined based on the equation (701). When the mode 2 is selected as the subject mode, the condition judgment based on the equation (702) is performed. When the mode 3 is selected as the subject mode, the condition judgment based on the equation (703) is performed. Note that the above formula (700), the formula (701), and the formula (700)

2) or when the calculation based on the conditional judgment based on the formula (703) is performed, the formula (700), the formula (701), the formula (702), and the formula (703) ), The luminance signal Y, the color signal IR-ΥI, and the color signal IΒ-ΥI are signals read from the frame memory 39. In addition, these expressions (700), (701), and (

A program for performing the case judgment based on the equation (702) or the equation (703) is described in advance in RA A4a of CP1J4.

Here, for example, if the above-described condition determination processing is performed on all pixel data stored in the frame memory 39, the number of processing becomes too large, and this condition is determined in real time. Judgment processing cannot be performed. However, in the present embodiment, since the above-described condition determination is performed only on the pixel data existing in the selected area, the CPU 4 can perform this condition determination processing in real time.

Accordingly, the CPU 4 is defined by the formula (700), (701), (7 () 2) or (7ϋ3) for each of the ginseng data in the selected area. The result of the determination as to whether the condition is met is obtained. Here, the fact that this pixel data power matches the condition defined by this equation (700), (701), (702) or (703) means that Picture This means that the color of the raw data matches and matches the color set as the target object.

Further, in step S304, the CPU 4 creates an object information table, which will be described later, in parallel with the condition determination processing, and stores it in a predetermined area of the RAM 4a. In this object information table, the coordinate of the object, which is set as the target object, indicates which line from which pixel position to which pixel position exists, and matches the color of the target object. And an object identification number indicating the number of the object having the desired color.

Next, with respect to the object information table, the _t> second 3 diagram described with reference to a second 3 figure is al, and object information te one table obtained from the example shown in the second 2 Figure is shown I have. The line position indicates the number of the line on which an object having the same color as the target object exists by the Y coordinate, and the start pixel position indicates the line indicated by this line position. In X, the coordinate position of the first pixel data of the holiday that has the same color as the target holiday is indicated by the X coordinate, and the end pixel location is defined as 1Ξ1 in the line indicated by this line location. The X-coordinate indicates the coordinate position of the last pixel data of the object having the same color as the target object, and the object identification IJ number is the number of objects recognized as objects having the same color as the target object. A number indicating whether the object is an eye.

For example, since the object される shown in FIG. 22 exists in 1 of the pixel 245 to pixel 246 of the line 161, the object information table contains the information よう as shown in FIG. Then, “16 1” is stored as the line identification, “2 45” is stored as the start pixel position, and “2 4 6” is stored as the end pixel position. Further, “1” is stored as the object identification number indicating the object Α. The same applies to the following 162-line to 190-line, and a description thereof will be omitted. Next, as shown in Fig. 23, in the 91 line, the object A exists between 221 pixels and 258 pixels, and object B exists between 318 pixels and 319 pixels. As shown in (1), as information indicating the object A, "191" is stored as the line ban number, "221" is stored as the start pixel position, and “258” is stored as the pixel position, and “1” is stored as the object identification number. Further, in the subject information table, “191” is stored as the line number as information indicating the object B, and “318” is stored as the start pixel position. Then, “63 19” is stored as the end pixel position, and “2” is stored as the object identification number. 6

In step S30, a minimum window including an object having the same color as the target object is set. In the example shown in Fig. 22, the minimum window containing the object Λ is defined as a window defined in the range of 2 16 ≤ X ≤ 27 3, 16 1 ≤ Y ≤ 20 2 Do W _A is set, and as the smallest window that includes vacation B, 3 09 ≤ X ≤

A window W _B defined in the range of 3 5 8, 1 9 1 Y ≤ 2 3 1 is set.

In step S306, m is initialized to the minimum object identification number stored in the physical information table. Here, m is a mere variable and has a value from the smallest object identification number stored in the object information table to the largest object identification number.

In step S307, the m-th pin set in step S305 based on the window center vector stored in the target object history table described later. It is determined whether or not the expected position coordinates described later exist in the do. This is to determine which of the object か and the object B is the target object-C.

First, with reference to FIG. 24, a description will be given with reference to the object leave history table. Fig. 24 shows an example of the target object history table FIG. The target object history table stores information on the coordinate position of the object determined to be the target object in each field. The field number is a temporary number that is reset every 30 fields, and is a sequential number that is sequentially assigned to each field. The window X coordinate is the

This is data indicating the range in the X-axis direction of the window set in 05 by X-coordinate values. What is the window Y coordinate?

This is data indicating the range in the Y-axis direction of the window set in 5 by the Y coordinate value. The window center vector (△ XΔγ) is defined as the window center position set in step S305, which is the center position of the imaging screen (X = 384, Y Y20). A vector indicating the direction and how far away from)

There is.

For example, in the example of the target object history table shown in FIG. 24, at the time indicated by the field number 17, the window range set for the target object is 3 1 2 X ≤ 3 62 and 1 86 ≤ Y ≤ 2 28 It also indicates that the center position of the window is displaced from the center position of the imaging screen in the direction and distance indicated by the window center vector (147 + 87). . The same applies to the window X sputum markers, window Υ coordinates, and window center vector created at the time indicated by field number 及び 8 and field number 19. Description is omitted.

Here, each of the field center 17 and the field number 18 and the field number 19 are respectively stored in the window center vectors stored at the respective time points. Looking at the data, there is no significant change in the values of the data representing these three window center vectors. This is not because the target object is not moving, but this window center vector This is because the torque indicates the movement vector of the target object from the previous field. In this embodiment, the CPU 4 controls the pan-Z tilt drive mechanism 16 so that the center position of the window indicating the moved target object is positioned at the center of the imaging screen every field. Is controlling. Therefore, by setting the window center vector as a vector indicating the direction and distance of deviation from the center of the imaging screen, the window center vector is shifted from the front field. It represents the movement vector of the target object.

Step S307 will be described again with reference to the target object history table described above. In step S307, first, the predicted position coordinates exist in the first window W _Λ (16 ^ ^ 27 3, 16) ≤ Y ≤ 20 2 corresponding to the object Λ. Judge whether to do. The predicted position coordinates are position coordinates obtained from the window center vector one field before, stored in the target object history table described above. For example, full I one field number is set at the time of〗 9 the windowing central base-vector (△) (, ₃ delta Y ₁₃₎ so is a base-vector (one 4 9 + 8 9), off Even if we look at the window center vector obtained at field number 20, we can expect it to be a vector near the vector (1 49 + 89) . Therefore, since the window center vector stored in the target object history table indicates the amount and direction of the coordinate deviation from the imaging center coordinates (384, 120), the window center vector is displayed. The window center position target set for the target object at the time of field No. 20 can be considered to be (335, 209). The center position coordinate of this window is the expected position coordinate.

Specifically, in step S307, the smallest window including the object A is 2 16 ≤ X ≤ 27 3 and

Target object history tape in the window W _A defined by the range The expected position coordinates (335, 209) obtained from the file do not exist. Therefore, the CPU 4 determines that the object A is not the set target object, and proceeds to step S308.

In step S308, m is incremented, and the process returns to step S307 again.

Returning to step S307 again, this time, as the smallest window including the object B, the range of 309 ≤ X ≤ 358 and 19.1 ≤ Y≤2 3 1 Some defined © fin de W _B, the expected position匾coordinates (3 3 5 2 0 9) determines whether there. Oite to this stearyl-up, in the windowing W _B, the expected position anonymous coordinates (3 3 5, 2 0 9

), CPU 4 determines that the object B is the set target object, and proceeds to step S309.

In step S309, the CPU 4 converts the coordinates of the window WB defined in the range of 3109 ≤ X ≤ 358 and 191 ≤ Y ≤ 231 into the target object of the RAM 4a. Field Acknowledgment in History Table 2

The window X coordinate and the ゥ^γ window Y coordinate are stored in the area indicated by 0. Also, the CPU 4 calculates the center coordinates of the window W _B from the coordinates of the window W _B , and uses the center coordinates of the window as the window center vector in the RAM 4 a. Remember. In the example described above, the center vector of the window is a solid

+ 9 1) is stored.

In step S310, the CPU 4 moves the center of the window W2 to the center of the imaging screen based on the window center vector newly stored in step S309. The tilt Z pan drive mechanism 16 is controlled so as to match. Specifically, the CPU 4 outputs a control signal to the motor drive circuit 16b based on the window center vector.

In step S311 CPU 4 is the window center vector The offset value is supplied to the evaluation value generation circuit 62 of the force control circuit 34 on the basis of the value. The offset value is an offset value supplied to each counter provided in the window pulse generation circuits 625 and 635 shown in FIGS. 3 and 4. When the offset value is not supplied to each of the window pulse generators 6 25 and 6 35, as shown in FIGS. 6A and 6B, each window The center coordinates of W1 to W11 coincide with the coordinates of the center of the imaging screen. However, when this offset value is supplied from the CPU 4 to the respective window counters of the window pulse generation circuits 625 and 6335, the counter value of each counter is based on this offset value. Changed. Therefore, the center coordinates of each window W1 to W11 are changed based on this offset value. When the offset value is supplied to the focus control circuit 34 in step S311, the process returns to step S100.

The present invention has the following effects.

The area search circuit 3 selects an area in which pixel data of the same color as the target object exists, and performs a condition determination process only on the pixel data existing in the selected area. The position of the target object can be grasped in real time without placing a processing burden on the CPU 4.

Also, the object mode is set according to the set color of the target object, and the condition judgment calculation by the area detection circuit 38 and the condition judgment by the CPU 4 are performed according to the set object mode. Since the calculation is variable, the subject can be accurately recognized regardless of the color of the set subject rest.

In addition, since all the condition determination processing performed in the area search circuit 38 is determined by the hardware circuit, all the pixel data supplied from the encoder 37 are processed in real time. Condition determination can be performed.

Further, even when there are a plurality of objects having the same color as the target object, an object information table having position information on each object and a target object history table having information on a movement history of the target object are provided. Since the target object is created, the target object can be accurately recognized.

In addition, since the position to which the target object has moved is calculated and the offset value based on the position is supplied to the window pulse generation circuit, the center deer of each window W 1 to W 11 is used. Changed to correspond to the object. Therefore, even if the target object moves, each window can be set accurately for the moved target object, and an accurate evaluation value can be obtained for the moved target object. Therefore, automatic focus control can be performed.

The present embodiment has the following effects. First, a plurality of evaluation values can be obtained by combining a plurality of filter coefficients with windows of a plurality of sizes, so that various types of subjects can be handled.

Also, weight data is added to the evaluation / generation circuit, and a total evaluation value is obtained based on a plurality of evaluation values and ffi data corresponding to the evaluation values. However, the accuracy of the finally obtained evaluation value is improved. When the accuracy of the evaluation value is improved, the evaluation value curve draws a clear parabola in the vicinity of the force point, so that the maximum point of the evaluation value can be determined quickly. Therefore, the operation of the auto-force becomes faster.

In addition, since an evaluation value determined to be inappropriate when calculating the overall evaluation 愤 is selected from a plurality of evaluation values and is not used, the accuracy of the evaluation value is further improved. For example, if a small window does not provide a good rating, use the rating corresponding to a larger window than the smaller window to combine the focus. Since it is made to focus, at least it is possible to focus on some kind of subject, and it is possible to prevent the auto-focus operation from continuing for a long time.

Furthermore, when deciding the moving direction in order to adjust the focus, a majority decision method using weight data is used for changes in a plurality of evaluation values, so that a small number of sample points and Small movements within the focal depth of the lens can judge right in the force direction.

When determining whether or not the maximum point of the evaluation value is the maximum point, the lens is moved from the maximum point to a distance of a predetermined multiple of the depth of focus, so that, for example, the peak of the evaluation value is flat. If so, it can be determined whether the lens is at its maximum when it moves a certain distance. Therefore, there is an effect that the focus point can be determined at high speed. For example, in order to determine whether the maximum point is the maximum point, it is necessary to prevent the focus from burning and the image signal to be large and fuzzy, thereby preventing an unnatural image from being output. it can.

When finding the maximum evaluation value at the force point, the evaluation state β matches the gap state of the overall evaluation value and the gap Z information stored in the R Λ GG, and the weight data Since the highest evaluation value of is selected as the maximum evaluation value, there is an effect that the value of the maximum evaluation value can be obtained accurately.

Claims

The scope of the claims

1. In an object recognizing device for recognizing a position of a target object, an image pickup means for outputting an electric image pickup signal;

The imaging screen formed by the imaging signal from the imaging unit is divided into a plurality of areas, and among the divided areas, pixel data having the same color component as the S-like subject exists. An error search means for selecting an error to be executed;

Storage means for storing all pixel data corresponding to an imaging signal from the imaging means;

Processing means for reading out pixel data corresponding to the area selected by the above-mentioned area search means from the storage means, and calculating a position of a target object based on the read-out pixel data; and An object recognizing device comprising:

2. In an imaging device for imaging a target subject,

Imaging means for outputting an electrical imaging signal;

A detection window is set for an imaging screen composed of the imaging signals obtained from the imaging means, and the evaluation value of the imaging signal in the set detection window is detected and evaluated. Focus control means for controlling a focus force based on the value;

Storage means for storing an image signal from the imaging means, corresponding to all pixel data;

The position of the target object is calculated based on the ^ data read from the storage means, and the focus is adjusted so that the calculated position of the target object matches the position of the detection window. An imaging apparatus comprising: processing means for controlling a control means.