CN101399924A - Automatic exposure method and device based on brightness histogram - Google Patents

Abstract

An automatic exposure method based on a brightness histogram, comprising the steps of: (1) dividing an input image into N*N image blocks; (2) calculating an average brightness; (3) calculating a statistical histogram; (4) dividing the brightness The histogram is divided into three areas; (5) Divide the underexposed area into m sub-areas (6) Divide the over-exposed area into n sub-areas (7) For N*N average brightness values, judge each average brightness value Whether it falls in each sub-area of the under-exposure area or the over-exposure area; (8) calculate the over-exposure value; (9) calculate the under-exposure value; (10) adjust the exposure time. An automatic exposure device for implementing the above automatic exposure method includes an imaging sensor, an image processing module, a histogram statistical module and an exposure calculation module. The invention can obtain faster and more accurate automatic exposure adjustment under different brightness and contrast conditions.

Description

Automatic explosion method and device based on brightness histogram

Technical field

The present invention relates to the automatic exposure control field of digital camera, particularly a kind of automatic explosion method and device based on brightness histogram.

Background technology

The major function of automatic exposure is to control the brightness of photographic images by the time for exposure of regulating digital camera imaging sensor (imaging sensor).

Automatic exposure is very big to the quality influence of output image.The target of automatic exposure is that the dark portion of the image details of taking is not lost, and highlights is overexposure not, and the suitable human eye of mean flow rate is watched simultaneously.

Most of traditional automatic explosion methods are searched for optimum exposure time and gain according to taking the photograph image by the sub-region right average brightness value.The shortcoming of this method is the image information over-simplification, poor effect when picture contrast is higher.

The step-length of automatic exposure search can influence the balance relation between search speed and the search stability.Time for exposure is adjusted the too small meeting of step-length increases searching times, causes reducing search speed; And if the time for exposure adjusts that step-length is long may to cause search to restrain, in the optimum exposure point unsettled phenomenon of vacillating now to the left, now to the right.Existing automatic explosion method majority can not be adjusted step-size in search dynamically according to the brightness situation of different scenes.

Summary of the invention

The present invention is the shortcoming that overcomes above-mentioned prior art, and purpose provides a kind of digital camera automatic explosion method and device based on brightness histogram.

For achieving the above object, technical scheme of the present invention is:

A kind of automatic explosion method based on brightness histogram is characterized in that this method comprises the steps:

(1) image processing module is divided into N*N image block with input picture;

(2) image processing module calculates the mean flow rate of each image block, and promptly the brightness value to all pixels in each image block is averaged, and obtains N*N average brightness value;

(3) image processing module calculates the statistic histogram of all N*N average brightness value;

(4) image processing module is divided into three zones with brightness histogram: under-exposure zone, region of normal exposure and overexposure zone, brightness value is under-exposure zone less than T1's, brightness value greater than T2 for the overexposure zone, brightness value is a region of normal exposure between T1 and T2;

(5) image processing module is divided into the m sub regions with under-exposure zone, distributes different weighted values for different subregions, and brightness is low more, and weight is big more;

(6) image processing module is divided into the n sub regions with the overexposure zone, distributes different weighted values for different subregions, and brightness is high more, and weight is big more;

(7) image processing module is for N*N average brightness value, judge whether each average brightness value drops in all subregion in under-exposure zone or overexposure zone, if certain average brightness value drops in certain under-exposure/overexposure subregion, then the count value with this subregion adds 1;

(8) the overexposure light value of statistics with histogram module computed image is the weighted sum value of the count value of each overexposure subregion, and weight is the weighted value that all subregion distributed:

$\mathrm{Over}\_\mathrm{value}=\underset{i=1~n}{\mathrm{\Σ}}W{o}_i*\mathrm{Overblock}\_{\mathrm{count}}_i$

Over_value is the overexposure light value in the formula, and Woi is that weight is calculated in the overexposure in i zone, and Overblock_counti is the overexposure count value in i zone;

(9) the under-exposure value of statistics with histogram module computed image is the weighted sum value of the count value of each under-exposure subregion, and weight is the weighted value that all subregion distributed:

$\mathrm{Under}\_\mathrm{value}=\underset{i=1~m}{\mathrm{\Σ}}W{u}_i*\mathrm{Underblock}\_{\mathrm{count}}_i$

Under_value is under-exposure value in the formula, and Wui is the under-exposure calculating weight in i zone, and Underblock_counti is the under-exposure count value in i zone;

(10) exposure calculating module is judged the direction of time for exposure adjustment and the step-length of adjustment according to overexposure light value Over_value and under-exposure value Under_value:

Step (10.1): whether at first judge Over_value greater than certain threshold value O_T1, if, think that then image is overexposure, need to reduce the time for exposure, change step (10.2); If not, change step (10.3);

Step (10.2): judge that further whether Over_value is greater than another threshold value O_T2, O_T2〉O_T1, if, then reduce the time for exposure with big step-length, otherwise, reduce the time for exposure so that small step is long, change step (10.5);

Step (10.3): if Over_value is less than O_T1, think that image does not have overexposure, judge further then whether image is under-exposure, the foundation of judging be Under_value greater than certain threshold value U_T1, Over_value is less than certain less threshold value O_T3, O_T3＜O_T1 simultaneously, if above condition is set up, think that then image is under-exposure, need to increase the time for exposure, change step (10.4); If above condition is false, change step (10.5);

Step (10.4): judge that further whether Under_value is greater than another threshold value U_T2, U_T2〉U_T1, if, then increase the time for exposure with big step-length, otherwise, with the small step long increase time for exposure;

Step (10.5):, think that then image to reach correct exposure, need not to adjust the time for exposure if overexposure and under exposed condition all are false;

Step (10.6): finish this adjustment.

A described N*N image block, N=8.

Described T1=80, T2=160.

Described Wu1=1.5, Wu2=1.2, Wu3=1.

Described Wo1=1, Wo2=1.2, Wo3=1.5.

Big step-length in the described step (10) is set to [(Exp-max-Exp_min)/4]/Exp_step, and little step-length is set to 2 ^*Exp_step, wherein Exp_min is the minimum exposure time of imaging sensor, and Exp_max is the maximum exposure time, and Exp_step is the minimum precision of adjusting the time for exposure.

A kind of for implementing the automatic exposure device of above-mentioned automatic explosion method, be characterized in: this automatic exposure device comprises imaging sensor, image processing module, statistics with histogram module and exposure calculating module, described imaging sensor outputs image to described image processing module, this image processing module is divided into N*N image block with this input picture, and to all pixels calculating mean flow rates in each image block, and this monochrome information exported to described statistics with histogram module, this statistics with histogram module is done statistics with histogram to the mean flow rate of each image block, and the statistics with histogram result exported to described exposure calculating module, this exposure calculating module judges the direction that exposure is adjusted and calculates the step-length of adjusting according to the histogram information of input, and should adjust direction and step-length is exported to described imaging sensor.

Compared with prior art, the invention has the beneficial effects as follows:

The distribution decision of step-length by brightness histogram adjusted in automatic exposure among the present invention, distance " object brightness zone " point far away more on brightness histogram, and its weight is big more.Such mechanism has guaranteed that exposure adjustment step-length changes with depth of exposure, reaches the balance between search speed and the search stability.

The present invention uses the distributed intelligence of brightness histogram can distinguish different exposure status, these situations be common luminance weighted mean value can't distinguish.So the present invention can reach more accurate brightness adjustment.

Histogram subregion thresholding among the present invention, histogram resolution, weight etc. are adjustable parameter, have improved system's flexibility ratio.

The present invention can adjust exposure value fast and accurately.It is more stable that the present invention regulates exposure value, can effectively avoid owing to do not restrain the explosure flash phenomenon that causes.

Description of drawings

Fig. 1 is the flow chart of automatic explosion method of the present invention.

Fig. 2 is that the inventive method is to histogrammic subregion exemplary plot.

Fig. 3 is the composition schematic diagram of automatic exposure device of the present invention.

Embodiment

Below in conjunction with accompanying drawing and embodiment the present invention is done to describe in further detail, but should not limit protection scope of the present invention with this.

See also Fig. 1 and Fig. 2.Accompanying drawing 1 is the flow chart of automatic explosion method of the present invention.Accompanying drawing 2 is that the inventive method is to histogrammic subregion exemplary plot.

1. at first, image processing module is divided into N*N image block with input picture, for example 8*8 image block.

2. image processing module calculates the mean flow rate of each image block, and promptly the brightness value to all pixels in each image block is averaged, and obtains N*N average brightness value.

3. image processing module calculates the statistic histogram of all N*N average brightness value.The histogrammic method of counting statistics belongs to the basic skills of signal processing, be summarized as follows: the image brightness excursion is divided into M gradient, for example for the imageing sensor of 8bit, its brightness excursion is 0～255, it is divided into M=16 gradient, and then histogrammic brightness resolution is 256/16=16.For N*N average brightness value, judge which rank of brightness step its brightness value belongs to, and then the count value with this brightness step adds one.Travel through all N*N brightness value to obtain the Luminance Distribution histogram information of all brightness values.

4. image processing module is divided into three zones with brightness histogram: under-exposure zone, region of normal exposure and overexposure zone, see accompanying drawing 2.Brightness value is under-exposure zone less than T1's, brightness value greater than T2 for the overexposure zone, brightness value is a region of normal exposure between T1 and T2.

For example: T1=80, T2=160.

5. image processing module is divided into the m sub regions with under-exposure zone, sees accompanying drawing 2.For example in the present embodiment, under-exposure zone is divided into 3 sub regions, is respectively: under1, under2, under3.For different subregions distribute different weighted values (be respectively Wu1, Wu2, Wu3), brightness is low more, weight is big more.For example in this realization example, Wu1〉Wu2〉Wu3.Wu1=1.5 for example, Wu2=1.2, Wu3=1.

6. similar, image processing module is divided into the n sub regions with the overexposure zone.For example in the present embodiment, the overexposure zone is divided into 3 sub regions, is respectively: over1, over2, over3.For different subregions distribute different weighted values (be respectively Wo1, Wo2, Wo3), brightness is high more, weight is big more.For example in this realization example, Wo1＜Wo2＜Wo3.Wo1=1 for example, Wo2=1.2, Wo3=1.5.Distribution to different overexposures zone can be obtained by experiment according to the characteristic of different sensors.

7. image processing module judges for N*N average brightness value whether each average brightness value drops in all subregion in under-exposure zone or overexposure zone.If certain average brightness value drops in certain under-exposure/overexposure subregion, then the count value with this subregion adds 1.

8. the overexposure light value of statistics with histogram module computed image is the weighted sum value of the count value of each overexposure subregion, and weight is the weighted value that all subregion distributed:

$\mathrm{Over}\_\mathrm{value}=\underset{i=1~n}{\mathrm{\Σ}}W{o}_i*\mathrm{Overblock}\_{\mathrm{count}}_i,$ N=3 in this example.

Over_value is the overexposure light value in the formula, and Woi is that weight is calculated in the overexposure in i zone, and Overblock_counti is the overexposure count value in i zone.

9. similar, the under-exposure value of statistics with histogram module computed image is the weighted sum value of the count value of each under-exposure subregion, and weight is the weighted value that all subregion distributed:

$\mathrm{Under}\_\mathrm{value}=\underset{i=1~m}{\mathrm{\Σ}}W{u}_i*\mathrm{Underblock}\_{\mathrm{count}}_i,$ M=3 in this example.

Under_value is under-exposure value in the formula, and Wui is the under-exposure calculating weight in i zone, and Underblock_counti is the under-exposure count value in i zone.

10. exposure calculating module is judged direction (increase or shorten) that the time for exposure is adjusted and the step-length of adjusting according to the overexposure light value Over_value and the under-exposure value Under_value of above calculating.For example in this example, whether at first judge Over_value greater than certain threshold value O_T1, if, think that then image is overexposure (image is bright excessively), need to reduce the time for exposure.Further judge Over_value whether greater than another threshold value O_T2 (O_T2〉O_T1), if, then reduce the time for exposure with big step-length, otherwise, with the small step long reduction time for exposure.On the other hand, if Over_value, thinks that image does not have overexposure less than O_T1, judge further then whether image is under-exposure.The foundation of judging be Under_value greater than certain threshold value U_T1, Over_value is less than certain less threshold value O_T3 (O_T3＜O_T1) simultaneously.The purpose that detects the Over_value value simultaneously is at the bigger situation of some contrast (dark partially as integral image, but some part is very bright).If above condition is set up, think that then image is under-exposure (image is dark excessively), need to increase the time for exposure.Further judge Under_value whether greater than another threshold value U_T2 (U_T2〉U_T1), if, then increase the time for exposure with big step-length, otherwise, with the small step long increase time for exposure.If overexposure and under exposed condition all are false, think that then image to reach correct exposure, need not to adjust the time for exposure.

Described big step-length is set to [(Exp_max-Exp_min)/4]/Exp_step, little step-length is set to 2*Exp_step, wherein Exp_min is the minimum exposure time of imaging sensor, and Exp_max is the maximum exposure time, and Exp_step is the minimum precision of adjusting the time for exposure.

The step-length that each above threshold value and exposure are adjusted is according to the different qualities of imaging sensor (imaging sensor) and different.

Fig. 3 is the composition schematic diagram of automatic exposure device of the present invention.As shown in Figure 3, this automatic exposure device comprises imaging sensor, image processing module, statistics with histogram module and exposure calculating module, described imaging sensor outputs image to described image processing module, this image processing module is divided into N*N image block with this input picture, and to all pixels calculating mean flow rates in each image block, and this monochrome information exported to described statistics with histogram module, this statistics with histogram module is done statistics with histogram to the mean flow rate of each image block, and the statistics with histogram result exported to described exposure calculating module, this exposure calculating module judges the direction that exposure is adjusted and calculates the step-length of adjusting according to the histogram information of input, and should adjust direction and step-length is exported to described imaging sensor.

Described imaging sensor (Imaging Sensor) can use for example OV9650 series of products of OmniVision technology Co., Ltd.Above-mentioned image processing module, statistics with histogram module, automatic exposure are calculated module and can be realized in built-in system software, are software interface between the module.

The present invention proposes a kind of automatic explosion method and device based on brightness histogram.Core concept of the present invention is to have utilized the adjustment of the histogram information realization automatic exposure of brightness.Different with most of traditional automatic explosion methods, the present invention simply is reduced to single weighted average with image luminance information, but obtain the distributed intelligence of image light and shade by the statistics brightness histogram, containing much information increases greatly, can obtain quicker, automatic exposure adjustment more accurately under different brightness and contrast's conditions.

The above is preferred embodiment of the present invention only, is not to be used for limiting practical range of the present invention.Be that all equivalences of doing according to the content of the present patent application claim change and modification, all should be technology category of the present invention.

Claims (7)

1, a kind of automatic explosion method based on brightness histogram is characterized in that this method comprises the steps:

(1) image processing module is divided into N*N image block with input picture;

(2) image processing module calculates the mean flow rate of each image block, and promptly the brightness value to all pixels in each image block is averaged, and obtains N*N average brightness value;

(3) image processing module calculates the statistic histogram of all N*N average brightness value;

(4) image processing module is divided into three zones with brightness histogram: under-exposure zone, region of normal exposure and overexposure zone, brightness value is under-exposure zone less than T1's, brightness value greater than T2 for the overexposure zone, brightness value is a region of normal exposure between T1 and T2;

(5) image processing module is divided into the m sub regions with under-exposure zone, distributes different weighted values for different subregions, and brightness is low more, and weight is big more;

(6) image processing module is divided into the n sub regions with the overexposure zone, distributes different weighted values for different subregions, and brightness is high more, and weight is big more;

(7) image processing module is for N*N average brightness value, judge whether each average brightness value drops in all subregion in under-exposure zone or overexposure zone, if certain average brightness value drops in certain under-exposure/overexposure subregion, then the count value with this subregion adds 1;

(8) the overexposure light value of statistics with histogram module computed image is the weighted sum value of the count value of each overexposure subregion, and weight is the weighted value that all subregion distributed:

$\mathrm{Over}\_\mathrm{value}=\underset{i=1~n}{\mathrm{\Σ}}{\mathrm{Wo}}_i*\mathrm{Overblock}\_{\mathrm{count}}_i$

Over_value is the overexposure light value in the formula, and Woi is that weight is calculated in the overexposure in i zone, and Overblock_counti is the overexposure count value in i zone;

(9) the under-exposure value of statistics with histogram module computed image is the weighted sum value of the count value of each under-exposure subregion, and weight is the weighted value that all subregion distributed:

$\mathrm{Under}\_\mathrm{value}=\underset{i=1~m}{\mathrm{\Σ}}{\mathrm{Wu}}_i*\mathrm{Undsrblock}\_{\mathrm{count}}_i$

Under_value is under-exposure value in the formula, and Wui is the under-exposure calculating weight in i zone, and Underblock_counti is the under-exposure count value in i zone;

(10) exposure calculating module is judged the direction of time for exposure adjustment and the step-length of adjustment according to overexposure light value Over_value and under-exposure value Under_value:

Step (10.1): whether at first judge Over_value greater than certain threshold value O_T1, if, think that then image is overexposure, need to reduce the time for exposure, change step (10.2); If not, change step (10.3);

Step (10.2): judge that further whether Over_value is greater than another threshold value O_T2, O_T2〉O_T1, if, then reduce the time for exposure with big step-length, otherwise, reduce the time for exposure so that small step is long, change step (10.5);

Step (10.3): if Over_value is less than O_T1, think that image does not have overexposure, judge further then whether image is under-exposure, the foundation of judging be Under_value greater than certain threshold value U_T1, Over_value is less than certain less threshold value O_T3, O_T3＜O_T1 simultaneously, if above condition is set up, think that then image is under-exposure, need to increase the time for exposure, change step (10.4); If above condition is false, change step (10.5);

Step (10.4): judge that further whether Under_value is greater than another threshold value U_T2, U_T2〉U_T1, if, then increase the time for exposure with big step-length, otherwise, with the small step long increase time for exposure;

Step (10.5):, think that then image to reach correct exposure, need not to adjust the time for exposure if overexposure and under exposed condition all are false;

Step (10.6): finish this adjustment.

2, the automatic explosion method based on brightness histogram according to claim 1 is characterized in that a described N*N image block, N=8.

3, the automatic explosion method based on brightness histogram according to claim 1 is characterized in that T1=80, T2=160.

4, the automatic explosion method based on brightness histogram according to claim 1 is characterized in that Wu1=1.5, Wu2=1.2, Wu3=1.

5, the automatic explosion method based on brightness histogram according to claim 1, its feature in

Wo1＝1，Wo2＝1.2，Wo3＝1.5。

6, the automatic explosion method based on brightness histogram according to claim 1, the big step-length of its feature in described step (10) is set to [(Exp_max-Exp_min)/4]/Exp_step, little step-length is set to 2*Exp_step, wherein Exp_min is the minimum exposure time of imaging sensor, Exp_max is the maximum exposure time, and Exp_step is the minimum precision of adjusting the time for exposure.

7, a kind of for implementing the automatic exposure device of the described automatic explosion method of claim 1, it is characterized in that this automatic exposure device comprises imaging sensor, image processing module, statistics with histogram module and exposure calculating module, described imaging sensor outputs image to described image processing module, this image processing module is divided into N*N image block with this input picture, and to all pixels calculating mean flow rates in each image block, and this monochrome information exported to described statistics with histogram module, this statistics with histogram module is done statistics with histogram to the mean flow rate of each image block, and the statistics with histogram result exported to described exposure calculating module, this exposure calculating module judges the direction that exposure is adjusted and calculates the step-length of adjusting according to the histogram information of input, and should adjust direction and step-length is exported to described imaging sensor.

Images