WO2013088657A1

WO2013088657A1 - Projecting projector device, optical anti-glare method, and optical anti-glare program

Info

Publication number: WO2013088657A1
Application number: PCT/JP2012/007623
Authority: WO
Inventors: 亮磨大網
Original assignee: 日本電気株式会社
Priority date: 2011-12-16
Filing date: 2012-11-28
Publication date: 2013-06-20

Abstract

In the present invention, a difference image generation means (82) generates a difference image from the difference values between the pixel value of each pixel in a post-correction captured image and the pixel value of each pixel in a corresponding projected image. A foreground extraction means (83) extracts a foreground region indicating a foreground portion from a binarized image, which is an image resulting from binarizing the difference image on the basis of a threshold. An anti-glare processing means (85) performs a glare-reducing process on each pixel of the projected image corresponding to a region containing a head. A dynamic threshold calculation means (86) extracts a region corresponding to the background region, which is the region other than the foreground region in the binarized image, from the projected image and from the post-correction captured image, and on the basis of the difference in pixel values of each corresponding pixel between the extracted images, calculates the threshold used in binarizing the difference image for each pixel.

Description

Projection type projection device, light anti-glare method, and light anti-glare program

The present invention relates to a projection type projection device, a light anti-glare method, and a light anti-glare program for reducing the glare of light projected when an image is projected.

When a presentation is performed using a projector that projects an image on a screen, the presenter has many opportunities to explain by standing in front of the projected image. In this case, since the light projected from the projector is directly irradiated to the presenter's eyes, the presenter often feels the light very dazzling. Therefore, various methods for projecting an image so as to alleviate the glare felt by the presenter are known.

Patent Document 1 describes a display device that controls the brightness of image light projected onto an area where a person is located. The display device described in Patent Document 1 captures a screen with a camera attached to a projector, and when a presenter is detected in the captured region, the brightness of the image light projected on the region where the presenter is located is reduced.

Patent Document 2 describes a projector that prevents various types of information included in a projected image from being displayed on the face of a presenter. The projector described in Patent Document 2 detects a human face area from a captured image obtained by photographing the projected image, and projects an image in which the area corresponding to the face area is replaced with a predetermined color.

Non-Patent Document 1 describes a projector system that adjusts light of a projector so that a speaker does not dazzle even when standing in front of the projector. The projector system described in Non-Patent Document 1 detects a speaker's face area from an image including a screen and the speaker, and displays a mask image superimposed on the detected area.

Non-Patent Document 2 describes a person detection method that improves the accuracy of detecting a person in a high illumination environment. In the method described in Non-Patent Document 2, brightness adjustment is performed between a projected image and a captured image obtained by capturing the projected image with a camera, and a difference image is created between the images after brightness adjustment, thereby Is detected.

Note that Patent Document 3 describes an image processing apparatus that extracts an object region without being affected by a pixel value of a background region. The image processing apparatus described in Patent Literature 3 obtains a difference value between input image data and background image data, and detects an obstacle candidate region using a predetermined binarization threshold.

JP 2000-305481 A JP 2010-34820 A JP 2000-57323 A

In order to reduce the glare of the presenter, it is effective to reduce the light projected on the presenter. On the other hand, the projected slide is an important material for the viewer. Therefore, it is desirable that the range in which the projected light is reduced is as close as possible to the user's head. Therefore, it is desirable to appropriately recognize the head of the presenter and reduce the light at that portion.

The display device described in Patent Document 1 detects a presenter by paying attention to a luminance change in which the lower part of the screen is darkened by the body and legs of the presenter in an image taken by a camera. However, since the display device described in Patent Document 1 does not detect a person, there is a problem that there are many false detections.

Also, in the system described in Non-Patent Document 1, there is a problem that face detection does not operate correctly and the anti-glare function does not work because the slide is superimposed on the presenter's face. Further, the system described in Non-Patent Document 1 has a problem that when a face is originally placed in a slide, even the face is detected and the anti-glare function is erroneously activated.

Patent Document 2 describes that a projector performs normalization including distortion correction, brightness, and brightness correction on a comparative image. However, since the specific correction method is not described in Patent Document 2, it is difficult to say that the projector described in Patent Document 2 can detect a person with high accuracy.

On the other hand, in Non-Patent Document 2, brightness adjustment is performed by performing processing for matching the histogram average value with the pixel value of the input image. However, as described in Non-Patent Document 2, when a predetermined luminance correction value is used, luminance correction is performed for a person to be extracted, and there is a case where a person cannot be appropriately extracted.

Therefore, when projecting an image by projecting light with high illuminance, the present invention reduces the glare of the light projected by detecting the person with high accuracy when a person is present in the image projection direction. It is an object of the present invention to provide a projection type projection apparatus, a light anti-glare method, and a light anti-glare program that can be reduced.

A projection type projection apparatus according to the present invention projects a projection image, which is an image projected from the projection means, and an image obtained by capturing the projection image or a foreground that is a moving object existing between the projection image and the projection means. Pixel value correction that corrects the pixel value of the pixel in the background that is a part other than the foreground in the captured image so that it approaches the pixel value of the pixel in the corresponding projection image. Means, a difference image generating means for generating a difference image from a difference value between a pixel value of each pixel in the captured image after correction and a pixel value of each pixel in the corresponding projection image, and a difference image based on a threshold value 2 A foreground extracting means for generating a binarized image that is a binarized image and extracting a foreground area indicating a foreground portion from the binarized image; and a moving object from an area corresponding to the foreground area in the photographed image A head detecting means for detecting the head, an anti-glare processing means for reducing glare on each pixel of the projection image corresponding to the area including the head, and an area other than the foreground area in the binarized image. An area corresponding to a certain background area is extracted from each of the projected image and the corrected captured image, and a threshold value used for binarization of the difference image based on a difference in pixel values at each corresponding pixel between the extracted images. And a foreground extraction unit that generates a binarized image from the difference image based on the threshold value calculated by the dynamic threshold value calculating unit, and from the binarized image, The foreground area is extracted, and the pixel value correcting means corrects the pixel values of the pixels other than the foreground area extracted by the foreground extracting means.

The anti-glare method according to the present invention projects a projection image, which is an image projected from the projection means, and an image obtained by photographing the projection image or a foreground, which is a moving object existing between the projection image and the projection means, And the pixel value of the pixel in the background that is a part other than the foreground in the photographed image is corrected so as to be close to the pixel value of the pixel in the corresponding projection image, and after the correction A difference image is generated from the difference value between the pixel value of each pixel in the captured image and the pixel value of each pixel in the corresponding projection image, and the binarized image is an image obtained by binarizing the difference image based on the threshold value A foreground region indicating the foreground part is extracted from the binarized image, the head of the moving object is detected from the region corresponding to the foreground region in the captured image, and the projection image corresponding to the region including the head For each pixel The processing corresponding to the background region in the binarized image is extracted from the projected image and the corrected captured image, and the corresponding pixels between the extracted images are processed. A threshold value used for binarization of the difference image is calculated for each pixel based on the difference between the pixel values in the image, and when extracting the foreground region, a binarized image is generated from the difference image based on the calculated threshold value. The foreground area is extracted from the binarized image, and the pixel values of the pixels other than the extracted foreground area are corrected.

An anti-glare program according to the present invention is a foreground that is a projected image that is an image projected from a projection unit on a computer and an image obtained by capturing the projection image or a moving object that exists between the projection image and the projection unit. Is compared with the captured image that is captured with the projected image, and the pixel value of the pixel in the background, which is a portion other than the foreground in the captured image, is corrected to be close to the pixel value of the pixel in the corresponding projected image. Pixel value correction processing, difference image generation processing for generating a difference image from the difference value between the pixel value of each pixel in the corrected captured image and the pixel value of each pixel in the corresponding projection image, and based on the threshold value of the difference image A foreground extraction process for generating a binarized image that is a binarized image and extracting a foreground region indicating a foreground portion from the binarized image, corresponding to a foreground region in a captured image Head detection processing for detecting the head of a moving object from the area, anti-glare processing for reducing glare in each pixel of the projection image corresponding to the area including the head, and areas other than the foreground area in the binarized image The region corresponding to the background region is extracted from the projected image and the corrected captured image, and is used for binarization of the difference image based on the difference in pixel value at each corresponding pixel between the extracted images. A dynamic threshold value calculation process for calculating a threshold value for each pixel is executed, and in the foreground extraction process, a binary image is generated from the difference image based on the threshold value calculated in the dynamic threshold value calculation process, and the binary image The foreground area is extracted from the image data, and the pixel value correction process corrects the pixel values of the pixels other than the foreground area extracted by the foreground extraction process.

According to the present invention, when projecting an image by projecting light with high illuminance, if there is a person in the projection direction of the image, the glare of the projected light is detected by detecting the person with high accuracy. Can be reduced.

It is a block diagram which shows the structural example of 1st Embodiment of the projection type projector by this invention. It is explanatory drawing which shows the example of a correspondence of the brightness | luminance of a camera image, and the brightness | luminance of a slide image. It is explanatory drawing which shows the example of the process which produces | generates a mask image. It is explanatory drawing which shows the example of an anti-glare process. It is a flowchart which shows the operation example of the projection type projector of 1st Embodiment. It is a flowchart which shows the other operation example of the projection type projector of 1st Embodiment. It is a block diagram which shows the structural example of 2nd Embodiment of the projection type projector by this invention. It is explanatory drawing which shows the example of the brightness | luminance difference between the camera image and slide image after brightness correction. It is a block diagram which shows the outline | summary of the projection type projector by this invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Embodiment 1. FIG.
FIG. 1 is a block diagram showing a configuration example of a first embodiment of a projection type projection apparatus according to the present invention. The projection type projection apparatus according to the first embodiment includes a geometric correction unit 101, a difference image generation unit 102, a foreground extraction unit 103, a head detection unit 104, an anti-glare processing unit 105, and a pixel value correction unit 121. A pixel value correction model calculation unit 122, a slide change determination unit 123, and a dynamic threshold calculation unit 124.

The geometric correction unit 101 inputs an image captured by the camera (hereinafter referred to as a camera image) and matches the shape of a slide image (hereinafter referred to as a slide image) projected on the screen by the projection type projector. Geometrically correct the camera image.

Specifically, when the projector is projecting a rectangular slide image, the geometric correction unit 101 geometrically corrects the camera image captured by the camera into a rectangle according to the shape of the slide image. Hereinafter, the camera image geometrically corrected by the geometric correction unit 101 may be referred to as a geometrically corrected image. As the geometric correction method, for example, a widely known general method such as a method using a projective transformation matrix is used.

Here, the camera image is an image of a slide image projected on the screen. In addition, when a presenter stands in front of a slide image projected on a screen for explanation, a presenter exists between the slide image and the projector. In this case, the camera image means an image obtained by photographing the presenter together with the slide image. In this embodiment, when it is described as a camera image, it means an image in which a slide image is taken or an image in which both a slide image and a presenter are taken.

In addition, since the presenter photographed as a camera image is an object that moves while projecting a slide image, the presenter may be referred to as a moving body in the following description. Further, from the viewpoint of the slide viewer, the slide video is sometimes referred to as the background, and the presenter present before the slide video is sometimes referred to as the foreground.

The pixel value correction model calculation unit 122 generates a model that defines a method for correcting the pixel value of the camera image so as to approach the pixel value of the corresponding slide image (hereinafter referred to as a pixel value correction model). A geometrically corrected image is used for this camera image. Note that a luminance value can also be used as the pixel value.

Generally, when a slide image is shot with a camera, the pixel value of the corresponding pixel differs between the shot image (that is, the camera image) and the slide image. Therefore, by correcting the pixel value of the pixel in the camera image to be close to the pixel value of the pixel in the slide image, it is possible to improve the accuracy of comparing the two images.

The pixel value correction model calculation unit 122 generates a pixel value correction model based on the relationship between the pixel value of the pixel in the slide image and the pixel value of the pixel in the camera image corresponding to each pixel of the slide image. However, in the present embodiment, the pixel value correction model calculating unit 122 compares the pixel value of the pixel in the background portion among the pixels in the camera image with the pixel value of the pixel in the corresponding slide image.

An appropriate pixel value correction model can be obtained by comparing pixel values of pixels only in a portion other than the foreground portion to be extracted (ie, background portion) except for a portion that does not correspond between the slide image and the camera image (ie, foreground portion). Can be generated. Note that the foreground area indicating the foreground portion of the camera image is specified by the foreground extraction means 103 described later. Hereinafter, the information indicating the foreground area may be referred to as foreground information.

The pixel value correction model calculation unit 122, for example, uses a least-square estimation method to calculate the correspondence between the pixel value of a pixel in the slide image and the pixel value of the pixel in the background portion of the camera image corresponding to each pixel of the slide image. The pixel value correction model may be generated from the correspondence relationship. Further, the pixel value correction model calculation unit 122 may generate a pixel value correction model based on a conversion process for matching the histograms of the images.

FIG. 2 is an explanatory diagram showing an example of the correspondence between the brightness of the camera image and the brightness of the slide image. In the graph illustrated in FIG. 2, the horizontal axis indicates the brightness of the camera image, and the vertical axis indicates the brightness of the slide image. Also, the black rhombus in FIG. 2 indicates the correspondence between the brightness of the camera image and the brightness of the slide image. Further, white squares in FIG. 2 indicate the result of specifying the correspondence between the pixel value of the pixel in the slide image and the pixel value of the pixel in the camera image using the least square estimation method. In FIG. 2, the correspondence between the two is represented in a linear form (linear).

Further, the pixel value correction model calculating unit 122 may generate one pixel value correction model for the entire image, or may generate a pixel value correction model for each partial region of the image. Further, when generating a pixel value correction model for each region, the pixel value correction model calculating unit 122 may separately generate a pixel value correction model for the vicinity of the boundary of the region. For example, there may be a case where the correspondence relationship between pixel value changes is different between the vicinity of the center of the projected image and the periphery. In such a case, by generating a pixel value correction model for each partial region of the image, it is possible to improve the accuracy of converting pixel values.

The pixel value correcting unit 121 compares the slide image and the camera image, and corrects the pixel value of the pixel in the background portion other than the foreground of the camera image. Specifically, the pixel value correction unit 121 makes the pixel value of the background of the camera image close to the pixel value of the corresponding slide image based on the pixel value correction model calculated by the pixel value correction model calculation unit 122. to correct. For example, when the pixel value correction model is expressed in a linear format, the pixel value correcting unit 121 corrects the pixel value by multiplying the pixel value of the background portion of the camera image by a certain constant and adding the certain value. Good. Further, the pixel value correction model may not be expressed in a line format. For example, an equation approximated by a polynomial or a broken line may be used as the pixel value correction model.

The pixel value correction unit 121 corrects the pixel value based on the pixel value correction model calculated by the pixel value correction model calculation unit 122 for the background portion. Therefore, it can be said that the pixel value correction unit 121 corrects the pixel values of the pixels other than the foreground area specified by the foreground extraction unit 103.

The difference image generation unit 102 calculates the difference between the pixel value of each pixel in the camera image after correcting the pixel value (hereinafter referred to as a pixel value corrected image) and the pixel value of each corresponding pixel in the slide image. A difference image is generated.

The foreground extraction unit 103 generates an image obtained by binarizing the difference image generated by the difference image generation unit 102 based on a threshold value determined for each pixel (hereinafter referred to as a binarized image). Then, the foreground extraction unit 103 extracts a foreground area indicating the foreground from the binarized image. This threshold value is dynamically determined by a dynamic threshold value calculation unit 124 described later.

The foreground extraction means 103 specifies, for example, a pixel in which the pixel value of each pixel in the difference image is greater than a threshold value. The foreground extraction unit 103 determines that the block area is a part of the foreground when the predetermined number of pixels are included in a predetermined block area. Then, the foreground extraction means 103 identifies a set of block areas determined as a part of the foreground as the foreground area. A portion other than the foreground area corresponds to the background. An image indicated by the foreground area (that is, a set of block areas) specified in this way may be referred to as a mask image.

The foreground extraction means 103 notifies the dynamic threshold value calculation means 124 and the pixel value correction model calculation means 122 of information indicating the extracted foreground area.

FIG. 3 is an explanatory diagram showing an example of processing for generating a mask image. When the camera image illustrated in FIG. 3A is input, the difference image generation unit 102 generates the difference image illustrated in FIG. The foreground extraction unit 103 generates a binarized image illustrated in FIG. 3C by comparing with a threshold value determined for each pixel. Further, the foreground extraction means 103 determines whether or not each block area of a predetermined unit is a part of the foreground, and generates a mask image illustrated in FIG.

The head detection means 104 detects the head of the moving object from the area corresponding to the foreground extracted by the foreground extraction means 103 in the camera image. For example, the head detection unit 104 may detect the head by pattern recognition using a template or a feature amount describing a head pattern viewed from various directions of 360 degrees. Since a method for detecting the head from an image including a person or the like is widely known, detailed description thereof is omitted here.

The anti-glare processing means 105 performs a process of reducing glare on each pixel of the slide image corresponding to the region including the head detected by the head detecting means 104 from the camera image. Specifically, the anti-glare processing unit 105 performs a process of reducing the luminance of each pixel of the slide image corresponding to the region including the head.

The anti-glare processing means 105 may, for example, reduce the luminance value of the pixels included in the head region uniformly (for example, the luminance value to 0), and a pattern such as an ellipse or a circle whose luminance gradually increases as the distance from the center increases. Luminance may be reduced by placing a figure over the head region. Moreover, the area | region which reduces a brightness | luminance does not need to correspond with a head area | region, The rectangular area | region near the head area | region center may be sufficient. In addition, when the head detection unit 104 detects the head direction, the anti-glare processing unit 105 may reduce the luminance of an area where there is a high possibility that eyes in front of the head are present.

FIG. 4 is an explanatory view showing an example of the anti-glare process. As a result of inputting the camera image illustrated in FIG. 4A, a mask image illustrated in FIG. 4B is generated. The head detection means 104 detects the head from the part corresponding to the foreground area in the camera image. FIG. 4C shows the detected head with a black frame. The anti-glare processing unit 105 reduces glare by, for example, superimposing a black rectangular image illustrated in FIG. 4D on the slide image.

The dynamic threshold value calculation means 124 calculates a threshold value that the foreground extraction means 103 uses for binarization of the difference image for each pixel. First, the dynamic threshold value calculation unit 124 extracts regions corresponding to the background from the slide image and the pixel value correction image, respectively. The dynamic threshold value calculation unit 124 calculates a threshold value for each pixel based on the difference between the pixel values of the corresponding pixels between the extracted images. At this time, the dynamic threshold value calculation unit 124 calculates the threshold value larger as the difference between the pixel values is larger.

The pixel value at the coordinates (x, y) of the slide image is I _slide (x, y), the pixel value of the camera image is I _cam (x, y), and a correction function derived from the luminance correction model (hereinafter, luminance correction function) ) Is f (I) (I is a luminance value), the threshold value Th (x, y) of each pixel is calculated by Equation 1 exemplified below.

Th (x, y) = α | f (I _cam (x, y)) − I _slide (x, y) |

Where α (α> 1) is a multiplier. For example, a value such as 1.5 is set for α. From this equation, for example, a portion with a large pixel value difference such as a character region has a large threshold value, and a portion with a small pixel value difference such as a flat region has a small threshold value. As described above, the foreground extraction unit 103 generates a binarized image from the difference image based on the threshold value calculated by the dynamic threshold value calculation unit 124. Therefore, an effect that the foreground can be more easily distinguished is obtained.

The slide change determination means 123 determines whether or not the slide image has changed. For example, the slide change determination unit 123 may determine whether or not the slide image has changed based on a difference between frames (for example, whether or not the difference is a predetermined value or more). Further, the slide change determination unit 123 may determine whether or not the slide image has changed using a shot division method for detecting a division point in the time direction in a television image.

Thus, when the slide image changes, the pixel value correction model also changes. Therefore, the pixel value correction model calculation unit 122 calculates the pixel value correction model again when the slide change determination unit 123 determines that the slide image has changed. It should be noted that when determining a change in an image, it is desirable to use a slide image having a stable luminance rather than a camera image.

Geometric correction means 101, difference image generation means 102, foreground extraction means 103, head detection means 104, anti-glare processing means 105, pixel value correction means 121, pixel value correction model calculation means 122, slide The change determination unit 123 and the dynamic threshold value calculation unit 124 are realized by a CPU of a computer that operates according to a program (person detection program).

For example, the program is stored in a storage unit (not shown) of the projection type projection apparatus, and the CPU reads the program, and in accordance with the program, the geometric correction unit 101, the difference image generation unit 102, the foreground extraction unit 103, the head The detection unit 104, the anti-glare processing unit 105, the pixel value correction unit 121, the pixel value correction model calculation unit 122, the slide change determination unit 123, and the dynamic threshold calculation unit 124 may be operated.

Further, the geometric correction unit 101, the difference image generation unit 102, the foreground extraction unit 103, the head detection unit 104, the anti-glare processing unit 105, the pixel value correction unit 121, and the pixel value correction model calculation unit 122 Each of the slide change determination unit 123 and the dynamic threshold calculation unit 124 may be realized by dedicated hardware.

Next, the operation of the projection type projection apparatus of this embodiment will be described. FIG. 5 is a flowchart showing an operation example of the projection type projection apparatus of the present embodiment. In the present embodiment, it is assumed that the moving object is not captured in the camera image in the initial state. In the initial state, it is assumed that a model assumed by the situation is set as the initial model for the pixel value correction model, and a value assumed by the situation is also set as the initial value for the threshold value.

The geometric correction unit 101 corrects the camera image (step S1). The pixel value correction unit 121 corrects the pixel value of the geometrically corrected image based on the pixel value correction model (step S2). In the initial state, since no moving object is captured in the camera image, the pixel value correcting unit 121 corrects the pixel value of the entire geometrically corrected image.

The difference image generation means 102 generates a difference image from the difference value between the pixel value of each pixel in the pixel value correction image and the pixel value of each corresponding pixel in the slide image (step S3). The foreground extraction means 103 generates a binarized image from the difference image and extracts the foreground area (step S4). In the initial state, no moving object is captured in the camera image. Therefore, no foreground area is extracted from the binarized image.

On the other hand, the dynamic threshold value calculation means 124 calculates a threshold value for each pixel based on the difference between the pixel values of the corresponding pixels of the slide image and the pixel value corrected image (step S5). In the initial state, no moving object is captured in the camera image. For this reason, the dynamic threshold value calculation unit 124 calculates a threshold value for each pixel for the entire image.

Further, the pixel value correction model calculation unit 122 generates a pixel value correction model based on the relationship between the pixel values of the corresponding pixels of the slide image and the camera image (step S6). In the initial state, no moving object is captured in the camera image. Therefore, the pixel value correction model calculation unit 122 generates a pixel value correction model for the pixel values of the entire image.

Next, processing when a moving object exists in front of a slide image will be described. FIG. 6 is a flowchart showing another example of the operation of the projection type projection apparatus of this embodiment. The process from the geometric correction unit 101 geometrically correcting the camera image until the difference image is generated from the pixel value corrected image and the slide image is the same as the process shown in steps S1 to S3 in FIG.

The foreground extraction means 103 generates a binarized image from the difference image and extracts the foreground area (step S4). Here, since the moving body is captured in the camera image, the foreground area is extracted from the binarized image.

The head detection means 104 detects the head of the moving body from the area corresponding to the foreground area extracted by the foreground extraction means 103 in the camera image (step S11). The anti-glare processing unit 105 performs a process of reducing glare on each pixel of the slide image corresponding to the region including the head detected by the head detection unit 104 from the camera image (step S12).

On the other hand, the dynamic threshold value calculation unit 124 compares pixels other than the foreground area of the slide image and the pixel value correction image, and calculates a threshold value for each pixel based on the difference between the pixel values of the corresponding pixels (step S13). .

The pixel value correction model calculation unit 122 compares pixels other than the foreground area of the slide image and the camera image, and generates a pixel value correction model based on the relationship between the pixel values of the corresponding pixels (step S14). . Thereafter, the pixel value correction unit 121 corrects the pixel value of the geometrically corrected image based on the pixel value correction model generated by the pixel value correction model calculation unit 122 in step S14.

As described above, according to the present embodiment, the pixel value correcting unit 121 compares the slide image with the camera-captured image, and determines the pixel value of the background portion of the camera image as the pixel of the corresponding slide image. Correct it so that it is close to the value. The difference image generation unit 102 generates a difference image from the difference value between the pixel value of each pixel in the corrected camera image and the pixel value of each pixel in the corresponding slide image.

The foreground extraction means 103 generates a binarized image from the difference image based on the threshold value, and extracts the foreground area from the binarized image. The head detecting means 104 detects the presenter's head from an area corresponding to the foreground area in the camera image. The anti-glare processing means 105 performs a process of reducing glare on each pixel of the slide image corresponding to the region including the head.

Dynamic threshold calculation means 106 extracts regions corresponding to the background region in the binarized image from the slide image and the corrected camera image, respectively. Then, the dynamic threshold value calculation unit 106 calculates a threshold value used for binarization of the difference image for each pixel based on the difference between the pixel values of the corresponding pixels between the extracted images.

At this time, the foreground extraction unit 103 generates a binarized image from the difference image based on the threshold value calculated by the dynamic threshold value calculation unit 106, and extracts a foreground region from the binarized image. Further, the pixel value correcting unit 121 corrects the pixel values of the pixels other than the foreground region extracted by the foreground extracting unit 103.

With the above configuration, when projecting an image by projecting light with high illuminance, if there is a person in the projection direction of the image, the glare of the light projected by detecting the person with high accuracy Can be reduced.

Embodiment 2. FIG.
FIG. 7 is a block diagram showing a configuration example of the second embodiment of the projection type projection apparatus according to the present invention. In addition, about the structure similar to 1st Embodiment, the code | symbol same as FIG. 1 is attached | subjected and description is abbreviate | omitted. The projection type projection apparatus according to the first embodiment includes a geometric correction unit 101, a difference image generation unit 102, a foreground extraction unit 103, a head detection unit 104, an anti-glare processing unit 105, and a pixel value correction unit 121. A pixel value correction model calculation unit 122, a slide change determination unit 123, a dynamic threshold calculation unit 124, and a foreground determination unit 125.

That is, the projection type projection apparatus of the second embodiment is different from the projection type projection apparatus of the first embodiment in that the foreground determination means 125 is further provided.

The foreground determination means 125 determines the validity of the foreground area extracted by the foreground extraction means 103, and if it determines that the foreground area is appropriate, information indicating the foreground area extracted by the foreground extraction means 103 is used as the dynamic threshold calculation means 124 and the pixel. The value correction model calculation means 122 is notified.

When a moving body is photographed in the camera image, the moving body generally exists as a lump in a part of the screen. Therefore, when the foreground area is distributed over the entire screen, a change in the illuminance of the slide, not the foreground, may be detected. In addition, when the illuminance of the slide changes, when the pixel values of the camera image are compared in time series, the difference in luminance value is generally small.

Further, when the dynamic threshold value calculation unit 124 calculates the threshold value of the camera image in which only the background area is captured, the threshold value tends to be kept low, so that the entire screen is determined to be the foreground area according to the illuminance of the slide. There is a possibility.

Therefore, the foreground determination means 125 determines the validity of the foreground area extracted by the foreground extraction means 103 based on the time-series pixel value change of the camera image and the distribution of the foreground area with respect to the entire camera image.

When the foreground determination unit 125 determines that the foreground region has been extracted due to a change in illuminance, the pixel value correction model calculation unit 122 regenerates the luminance correction model, and the dynamic threshold calculation unit 124 Recalculate. Then, after the foreground extraction means 103 regenerates the mask image again, the foreground determination means 125 performs the above determination again.

In other words, the dynamic threshold value calculation means 124 extracts the area corresponding to the background area from the slide image and the corrected camera image using the foreground area determined to be appropriate by the foreground determination means 125. Further, the pixel value correction model calculation unit 122 generates a pixel value correction model.

In this way, by performing the determination process assuming that the difference between the slide image and the camera image after luminance correction is large, it is erroneously determined as the foreground when the luminance is changed only by the change in the illuminance of the projector. Can be prevented.

Hereinafter, a method by which the foreground determination unit 125 determines the foreground area will be specifically described. In the present embodiment, the foreground determination unit 125 first expands the mask image and performs determination on the expanded image. The foreground determination unit 125 performs determination by assuming, for example, an area obtained by expanding the mask image by a certain number of pixels as a foreground area. For example, in order to expand N pixels in the vertical and horizontal directions, the foreground determination unit 125 may perform processing by applying a maximum value filter having a window of (N + 1) × (N + 1) pixels.

Even if the foreground does not exist due to a change in the illuminance of the projector, the brightness difference between the pixel value of the camera image after brightness correction and the slide image may become large. Therefore, the foreground determination means 125 detects this state separately.

FIG. 8 is an explanatory diagram showing an example of the luminance difference between the camera image after the luminance correction and the slide image. As shown in FIG. 8A, when the difference between the slide image of N frames and the camera image is small, the number of pixels exceeding the threshold is small as shown in FIG. Here, as shown in FIG. 8C, it is assumed that the content of the slide image does not change, but the illuminance of the projector changes between the N frame and the N + 1 frame. In this case, as a result of the overall change in the luminance value of the camera image, a large number of pixels exceeding the threshold are extracted as shown in FIG.

Hereinafter, an example of determination processing performed assuming that the difference between the slide image and the camera image after the brightness correction is large will be described. In this determination processing, the feature that the luminance change between frames is not so large and the feature that the luminance change appears on the entire screen are used. Specifically, the foreground determination unit 125 compares the binarized images between two time-series frames, and the number of pixels exceeding the threshold (that is, the portion determined to be the background region) increases as a whole. Determine whether or not.

First, the foreground determination means 125 compares the binarized images between two frames. The foreground determination means 125 obtains the difference between the pixel value of the pixel of the target image and the luminance value of the pixel of the image one frame before the pixel exceeding the threshold. When this difference is within a certain threshold, the foreground determination means 125 checks the degree of variation in the position of the pixel. The foreground determination means 125 can determine the degree of variation by obtaining the variance of coordinate values. When the variance is equal to or greater than a certain value, it is considered that the image has spread over the entire screen. Therefore, the foreground determination unit 125 determines that a change in luminance due to a change in illuminance has occurred.

Note that the foreground determination means 125 may perform the determination process using the following method. In the determination process described below, the feature that the absolute value of the luminance change value between images to be compared is not so large and the feature that the luminance change appears on the entire screen are used.

Specifically, the foreground determination means 125 checks how much the difference between the pixel values exceeding the threshold at the time of binarization exceeds the threshold. When the difference from the threshold is within a certain threshold, the foreground determination means 125 checks the degree of variation in the position of the pixel. The foreground determination means 125 can determine the degree of variation by obtaining the variance of coordinate values. When the variance is equal to or greater than a certain value, it is considered that the image has spread over the entire screen. Therefore, the foreground determination unit 125 determines that a change in luminance due to a change in illuminance has occurred.

As described above, in this embodiment, the foreground determination means 125 determines the validity of the foreground area. Therefore, in addition to the effects of the first embodiment, the accuracy of determining the foreground region can be improved.

Next, the outline of the present invention will be described. FIG. 9 is a block diagram showing an outline of a projection type projection apparatus according to the present invention. A projection type projection apparatus according to the present invention includes a projection image (for example, a slide image) that is an image projected from a projection unit (for example, a projector) and an image obtained by capturing the projection image or between the projection image and the projection unit. A foreground (e.g., presenter, person) that is an existing moving body is compared with a captured image (e.g., a camera image) that is an image obtained by capturing the projected image together with the projected image. A pixel value correcting unit 81 (for example, a pixel value correcting unit 121) that corrects the pixel value of the pixel so as to approach the pixel value of the pixel in the corresponding projection image, and the pixel value of each pixel in the corrected captured image; Difference image generation means 82 (for example, difference image generation means 102) that generates a difference image from a difference value between the pixel value of each pixel in the corresponding projection image, and a difference A foreground extraction unit 83 (for example, foreground) that generates a binarized image that is an image obtained by binarizing the image based on a threshold and extracts a foreground region (for example, a mask image) indicating a foreground portion from the binarized image. Extraction means 103), head detection means 84 (for example, head detection means 104) for detecting the head of the moving body from the area corresponding to the foreground area in the captured image, and projection corresponding to the area including the head. Anti-glare processing means 85 (for example, anti-glare processing means 105) that performs processing to reduce glare on each pixel of the image, and a region corresponding to the background area other than the foreground area in the binarized image are projected Dynamic threshold calculation that extracts a threshold value used for binarization of a difference image for each pixel based on a difference between pixel values in each corresponding pixel between the extracted images and the captured image after correction Means 86 ( In example, a dynamic threshold calculation means) and.

The foreground extraction means 83 generates a binary image from the difference image based on the threshold value calculated by the dynamic threshold value calculation means 86, and extracts the foreground region from the binary image. The pixel value correcting unit 81 corrects the pixel values of the pixels other than the foreground area extracted by the foreground extracting unit 83.

With such a configuration, when projecting an image by projecting light with high illuminance, if there is a person in the projection direction of the image, the glare of the light projected by detecting the person with high accuracy can be reduced. Can be reduced.

In addition, the projection type projection apparatus uses the pixel value of the projected image corresponding to the pixel value of the captured image based on the relationship between the pixel value of the pixel of the projected image and the pixel value of the background pixel among the pixels of the corresponding captured image. You may provide the pixel value correction model calculation means (for example, pixel value correction model calculation means 122) which calculates the pixel value correction model which is a model which prescribed | regulated the method of correct | amending so that it may approximate to a value. Then, the pixel value correction model calculation unit may specify the background from the foreground region extracted by the foreground extraction unit 83, and the pixel value correction unit 81 may correct the pixel value of the captured image based on the pixel value correction model. .

Furthermore, the projection type projector may include a change determination unit (for example, a slide change determination unit 123) that determines whether or not the projection image has changed. The pixel value correction model calculating unit may generate a pixel value correction model when it is determined that the projection image has changed. In this case, the pixel value correction model calculation means can update the pixel value correction model at an appropriate timing.

In addition, the projection type projection apparatus determines the validity of the foreground region extracted by the foreground extraction unit 83 based on the time-series pixel value change of the captured image and the distribution of the foreground region with respect to the entire captured image (for example, , Foreground determination means 125) may be provided. Then, the dynamic threshold value calculation means 86 extracts areas corresponding to the background area from the projected image and the corrected captured image using the foreground area determined to be appropriate by the foreground determination means, and the pixel value correction model calculation means The pixel value correction model may be calculated by specifying the background of the captured image using the foreground area determined to be appropriate by the foreground determination means. In this case, the accuracy of determining the foreground area can be improved.

Further, the projection type projection apparatus may include a geometric correction unit (for example, a geometric conversion unit 101) that geometrically corrects the captured image in accordance with the shape (for example, a rectangle) of the projection image. Then, the pixel value correcting means may correct the pixel value of the captured image by comparing the projected image with the geometrically corrected captured image.

Further, the anti-glare processing means 85 may perform processing for reducing the luminance of each pixel of the projection image corresponding to the region including the head (for example, processing for superimposing a black image).

Further, the dynamic threshold value calculation means 86 may calculate the threshold value larger as the difference between the pixel values is larger (for example, the threshold value is calculated using the above equation 1).

As mentioned above, although this invention was demonstrated with reference to embodiment and an Example, this invention is not limited to the said embodiment and Example. Various changes that can be understood by those skilled in the art can be made to the configuration and details of the present invention within the scope of the present invention.

This application claims priority based on Japanese Patent Application No. 2011-275814 filed on Dec. 16, 2011, the entire disclosure of which is incorporated herein.

The present invention is preferably applied to a projection type projection apparatus that reduces glare of light projected when an image is projected.

DESCRIPTION OF SYMBOLS 101 Geometric correction means 102 Difference image generation means 103 Foreground extraction means 104 Head detection means 105 Anti-glare processing means 121 Pixel value correction means 122 Pixel value correction model calculation means 123 Slide change determination means 124 Dynamic threshold value calculation means 125 Foreground determination means

Claims

A projected image that is an image projected from the projection unit, and a captured image that is an image obtained by capturing the image obtained by capturing the projection image or a foreground that is a moving object existing between the projection image and the projection unit together with the projection image And a pixel value correction unit that corrects the pixel value of the pixel in the background that is a part other than the foreground in the photographed image so as to approach the pixel value of the pixel in the corresponding projection image;
Difference image generation means for generating a difference image from a difference value between a pixel value of each pixel in the captured image after correction and a pixel value of each pixel in the corresponding projection image;
Foreground extraction means for generating a binarized image that is an image obtained by binarizing the difference image based on a threshold, and extracting a foreground region indicating the foreground portion from the binarized image;
Head detection means for detecting the head of the moving body from an area corresponding to the foreground area in the captured image;
Anti-glare processing means for performing processing to reduce glare on each pixel of the projection image corresponding to the region including the head; and
An area corresponding to a background area that is an area other than the foreground area in the binarized image is extracted from the projected image and the corrected photographed image, and the pixel value of each corresponding pixel between the extracted images is extracted. Dynamic threshold calculation means for calculating a threshold value used for binarization of the difference image for each pixel based on the difference,
The foreground extraction unit generates a binarized image from the difference image based on the threshold calculated by the dynamic threshold calculation unit, extracts a foreground region from the binarized image,
The projection type projection apparatus, wherein the pixel value correction unit corrects pixel values of pixels in a portion other than the foreground region extracted by the foreground extraction unit.
A method for correcting a pixel value of a photographed image to be close to a pixel value of a corresponding projection image based on a relationship between a pixel value of a pixel of the projection image and a pixel value of a background pixel among the pixels of the corresponding photographed image A pixel value correction model calculating means for calculating a pixel value correction model that is a model that defines
The pixel value correction model calculation means specifies a background from the foreground region extracted by the foreground extraction means,
The projection type projection apparatus according to claim 1, wherein the pixel value correction unit corrects a pixel value of the captured image based on the pixel value correction model.
A change determining means for determining whether or not the projected image has changed,
The projection type projection apparatus according to claim 2, wherein the pixel value correction model calculation unit generates a pixel value correction model when it is determined that the projection image has changed.
Foreground determination means for determining the validity of the foreground area extracted by the foreground extraction means based on the time-series pixel value change of the captured image and the distribution of the foreground area with respect to the entire captured image,
The dynamic threshold value calculation means extracts the area corresponding to the background area from the projected image and the corrected captured image using the foreground area determined to be appropriate by the foreground determination means,
4. The projection type projection according to claim 2, wherein the pixel value correction model calculation unit calculates a pixel value correction model by specifying a background of the photographed image using the foreground region determined to be appropriate by the foreground determination unit. apparatus.
Comprising geometric correction means for geometrically correcting the captured image in accordance with the shape of the projected image;
5. The projection type projection device according to claim 1, wherein the pixel value correction unit corrects the pixel value of the captured image by comparing the projected image with the geometrically corrected captured image. 6.
The projection type projector according to any one of claims 1 to 5, wherein the anti-glare processing means performs a process of reducing the luminance of each pixel of the projection image corresponding to the region including the head.
The projection type projection apparatus according to any one of claims 1 to 6, wherein the dynamic threshold value calculation unit calculates the threshold value as the pixel value difference increases.
A projected image that is an image projected from the projection unit, and a captured image that is an image obtained by capturing the image obtained by capturing the projection image or a foreground that is a moving object existing between the projection image and the projection unit together with the projection image And correcting the pixel value of the pixel in the background that is a part other than the foreground in the captured image so as to approach the pixel value of the pixel in the corresponding projection image,
A difference image is generated from a difference value between a pixel value of each pixel in the captured image after correction and a pixel value of each pixel in the corresponding projection image,
Generating a binarized image that is an image obtained by binarizing the difference image based on a threshold, and extracting a foreground region indicating the foreground portion from the binarized image;
Detecting the head of the moving body from an area corresponding to the foreground area in the captured image;
Performing a process of reducing glare on each pixel of the projected image corresponding to the region including the head;
An area corresponding to a background area that is an area other than the foreground area in the binarized image is extracted from the projected image and the corrected photographed image, and the pixel value of each corresponding pixel between the extracted images is extracted. Based on the difference, a threshold value used for binarization of the difference image is calculated for each pixel,
When extracting a foreground region, a binarized image is generated from the difference image based on the calculated threshold value, and a foreground region is extracted from the binarized image,
A light anti-glare method characterized by correcting pixel values of pixels other than the extracted foreground region.
A method for correcting a pixel value of a photographed image to be close to a pixel value of a corresponding projection image based on a relationship between a pixel value of a pixel of the projection image and a pixel value of a background pixel among the pixels of the corresponding photographed image A pixel value correction model calculating means for calculating a pixel value correction model that is a model that defines
When calculating the pixel value correction model, the background is specified based on the foreground region extracted from the binarized image,
The light anti-glare method according to claim 8, wherein a pixel value of a captured image is corrected based on the pixel value correction model.
On the computer,
A projected image that is an image projected from the projection unit, and a captured image that is an image obtained by capturing the image obtained by capturing the projection image or a foreground that is a moving object existing between the projection image and the projection unit together with the projection image And a pixel value correction process for correcting the pixel value of the pixel in the background that is a part other than the foreground in the captured image so as to approach the pixel value of the pixel in the corresponding projection image,
A difference image generation process for generating a difference image from a difference value between a pixel value of each pixel in the captured image after correction and a pixel value of each pixel in the corresponding projection image;
A foreground extraction process for generating a binarized image that is an image obtained by binarizing the difference image based on a threshold value, and extracting a foreground region indicating the foreground part from the binarized image;
A head detection process for detecting the head of the moving body from an area corresponding to the foreground area in the captured image;
Anti-glare processing for reducing glare in each pixel of the projected image corresponding to the region including the head, and
An area corresponding to a background area that is an area other than the foreground area in the binarized image is extracted from the projected image and the corrected photographed image, and the pixel value of each corresponding pixel between the extracted images is extracted. Based on the difference, a dynamic threshold value calculation process for calculating a threshold value used for binarization of the difference image for each pixel is executed,
In the foreground extraction process, a binary image is generated from the difference image based on the threshold value calculated in the dynamic threshold value calculation process, and a foreground region is extracted from the binary image,
An anti-glare program for correcting pixel values of pixels in a portion other than the foreground area extracted by the foreground extraction process in the pixel value correction process.
On the computer,
A method for correcting a pixel value of a photographed image to be close to a pixel value of a corresponding projection image based on a relationship between a pixel value of a pixel of the projection image and a pixel value of a background pixel among the pixels of the corresponding photographed image A pixel value correction model calculation process for calculating a pixel value correction model that is a model that defines
In the pixel value correction model calculation process, the background is specified from the foreground region extracted by the foreground extraction process,
The program for anti-glare according to claim 10, wherein the pixel value of the photographed image is corrected based on the pixel value correction model in a pixel value correction process.