CN103366345A - Degraded video image restoration technology - Google Patents
Degraded video image restoration technology Download PDFInfo
- Publication number
- CN103366345A CN103366345A CN2012100940319A CN201210094031A CN103366345A CN 103366345 A CN103366345 A CN 103366345A CN 2012100940319 A CN2012100940319 A CN 2012100940319A CN 201210094031 A CN201210094031 A CN 201210094031A CN 103366345 A CN103366345 A CN 103366345A
- Authority
- CN
- China
- Prior art keywords
- image
- view
- video image
- weak signal
- video
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention (a degraded video image restoration technology) belongs to the field of video image processing and provides a method for restoring degraded video images based on a weak signal extraction method. Specifically, the method involves: forming an initial perspective view by extracting weak signals from the three primary colors of the video images; then using the grey-scale map of original video images to modulate the initial perspective view through a high-efficient guiding type image filter, forming a fine perspective view having edge details and further obtaining image transmissivity; at the same time obtaining a parameter called ''atmosphere light'' through statistics of the overall initial perspective view; and finally restoring restoration images with quite clear visual details and brilliant color by de-aliasing operation processing of the image transmissivity, the ''atmosphere light'' and the original images. The technology provided by the invention is mainly used in video monitoring industries for enhancing video definition and improving visual effects.
Description
Technical field
The invention belongs to the video image field, relate to a kind of method that the video image of degenerating is restored, be applied to video image processing, video image demonstration and transmission of video images.
Background technology
Video camera is under good weather condition, and the picture of shooting is clear bright, but under the grey haze in city, severe weather conditions such as misty rain, Sand-Dust Storm in Northern in the mountain or the picture of taking in the darker environment of light will thicken unclear.We call the video image that obtains in this case the Degenerate Graphs of original picture rich in detail.For the original picture rich in detail that obtains to restore, people have proposed many image processing techniquess, according to the course of technical development, always have 3 generation techniques;
The first generation: histogram equalization techniques
1) central idea of this technical finesse is from expanding to the uniformly distributing in whole tonal ranges between certain gray area of relatively concentrating the grey level histogram of original image.Histogram equalization carries out Nonlinear extension to image exactly, redistributes image pixel value, makes the pixel quantity in certain tonal range roughly the same; In other words, histogram equalization is exactly the histogram distribution of Given Graph picture to be changed over even distribution histogram distribute.This method is commonly used to increase the local contrast of many images, especially when the contrast of the useful data of image quite approaches.By this method, brightness can distribute at histogram better.
2) this technological merit is, all too bright or too dark image is very useful for background and prospect, such as can bring in the x-ray image better skeletal structure to show and over-exposed or under-exposed photo in better details.Be a quite intuitively technology and be inverse operation but a main advantage of this method is it, if known balanced function so just can recover original histogram, and calculated amount is also little.
3) but this technology has very large shortcoming, the gray level that is exactly image after the conversion reduces, some details disappears; Some image has the peak such as histogram, the after treatment factitious undue enhancing of contrast; It is indiscriminate to the data of processing, and it may increase the contrast of background noise and reduce the contrast of useful signal.
The second generation: dynamic contrast stretches
1) contrast stretching is the most basic a kind of greyscale transformation, what use is the simplest piecewise linear transform function, its main thought is the dynamic range of gray level when improving the image processing, be applicable to the processing of soft image, generally formed by two basic operations: statistics with histogram, determine image is carried out two flex points that gray scale stretches; Greyscale transformation,
The piecewise linear transform function of determining according to statistics with histogram carries out the mapping of grey scale pixel value.
2) this technological merit is to realize simply;
3) but this technology has very large shortcoming, can't process the very serious image of degeneration, also have the degraded image under the close shot condition to process to existing distant view in the image.
The third generation: homomorphic filtering
1) homomorphic filtering is a kind of image processing method that frequency is filtered and greyscale transformation combines, and it relies on the illumination/Reflectivity Model of image as the basis of frequency domain processing, utilizes the compression brightness range and strengthens the quality that contrast is improved image.The ultimate principle of homomorphic filtering is grey scale pixel value to be regarded as the product of illumination and two components of reflectivity.Because it is very little that illumination changes relatively, can be regarded as the low frequency composition of image, reflectivity then is the high frequency composition.By processing respectively illumination and reflectivity to the impact of grey scale pixel value, reach the purpose that discloses the shadow region minutia.
2) contrast of this technological merit image, color are all much improved.
3) but shortcoming also clearly, be exactly that the image noise is very serious, ring of light effect is arranged, its calculation of complex simultaneously, cost is expensive.
Summary of the invention
Be to solve the variety of problems that prior art exists, we have proposed a cover based on the weak signal extraction method through Long-Term Scientific Study and engineering practice, a kind of method that the video image of degenerating is restored.It is theoretical that its basic thought is based on a kind of priori that a large amount of observations and experiment summarizes go out: be exactly under the outdoor scene of good weather condition, in three passages of the RGB of the image block of a N*N size of camera picture federation at one close to 0 basic channel unit; When weather conditions become abominable (such as the grey haze in city, misty rain, the Sand-Dust Storm in Northern in the mountain), this basic channel unit value also can significantly increase, and the perspective rate of its added value and this image block is linear.
It comprises following step,
1) input video RGB picture signal at first;
2) by the extraction to weak signal in the video image three primary colors, form an initial perspective view;
3) simultaneously, by being added up, the initial perspective view overall situation obtains a parameter that is called " atmosphere light ";
4) again by an efficient guidance type image filter, the gray-scale map modulation initial perspective view with raw video image forms the meticulous skeleton view with edge details, and then obtains the image transmissivity;
5) use at last image transmissivity, " atmosphere light " to carry out the anti-aliasing calculation process with former figure, restore the palinspastic map that visually details is very clear, color is very bright-coloured.
6) if the brightness average of input picture is higher than a threshold value, according to flow processing as mentioned above; Otherwise at the input end of image, to the image negate, then pass through after the above-mentioned flow process, again the image negate; Can obtain like this low-light (level) hypograph brightness and contrast promotes.
Description of drawings
The system construction drawing of accompanying drawing 1. embodiment of the invention
The building-block of logic of the weak signal extractor module of accompanying drawing 2. embodiment of the invention
The building-block of logic of the rough skeleton view extractor module of accompanying drawing 3. embodiment of the invention
The building-block of logic of the atmosphere light extraction module of accompanying drawing 4. embodiment of the invention
The building-block of logic of the image transmissivity generator module of accompanying drawing 5. embodiment of the invention
The building-block of logic of the anti-aliasing module of accompanying drawing 6. embodiment of the invention
The building-block of logic of the overall brightness average statistical module of accompanying drawing 7. embodiment of the invention
Embodiment
Below in conjunction with the drawings and specific embodiments the present invention is done further and to specify.Be system construction drawing of the present invention as shown in Figure 1.As shown in Figure 1, the present embodiment is a kind of restoration disposal system of degraded image, is used for video camera because misty rain, grey haze, sand and dim and cause the picture signal of degenerating are repaired automatically.Described system comprises: by weak signal extractor module, rough skeleton view extractor module, atmosphere light extractor module, meticulous skeleton view generator module, anti-aliasing module frame chart and overall brightness average statistical module.
We know, according to image three primary colors (RGB) theory, a sub-picture is comprised of the pixel of M row and N row, and each pixel has again three passages of RGB to consist of, and in the picked-up of image, procedure for displaying, image is to transmit according to the mode of pixel stream.As shown in Figure 2, weak signal extractor module.Its Main Function is choosing minimum value in the R of each pixel in the image, G, three passages of B, and mathematical expression is: min (r, g, b), calculate one by one according to pixel stream.The data that enter, our called after RGB, the data of this module output, our called after RGB_min, the pixel map that its forms is called RGB_min figure.Data stream enters by name rough skeleton view extractor module as shown in Figure 3 subsequently, the FIFO array that this module is comprised of the capable pixel stream of M, select the moving window of a M*M size as a computing block, from first pixel of the first row, by pixel successively from left to right until last pixel of this row then be transferred to from first pixel of the second row.。。To the last last pixel of delegation moves the minimum value of choosing all data in this moving window at every turn, and mathematical expression is: min (local window); The physical significance of this module is in order to find image in the perspective rate of this point.Because the image of output is the image of similar mosaic sample, has lost edge details, so be called rough skeleton view.
As shown in Figure 4, the effect of this module is to generate " atmosphere light " parameter.This module is comprised of 3 submodules: totalizer, image blur histogram memory and atmosphere light optimum value finder.The image blur histogram memory is to add up video image each brightness in a two field picture always to have how many pixels, and storer is divided into 0～255 address space according to the brightness of image, and the statistical value of each brightness stores corresponding address into.Concrete is operating as: first pulse that enters effective district when the field signal that detects video image, initialization totalizer and image blur histogram memory are full 0, enter data effectively behind the district when video image, the pixel of each input, totalizer finds the number that once adds up before this brightness according to the brightness value of input from the memory location of correspondence, then add 1, then continue the next pixel of input, so circulation has obtained the statistic histogram of the Luminance Distribution of RGB_win_min figure this moment until the field signal of video image enters blanking zone in storer; Then open atmosphere light optimum value finder, it begins to read from the superlatively location of storer, and when the aggregate-value in the read memory was not 0, the value of this moment was maximal value, as " atmosphere light " A.
As shown in Figure 5, the effect of this module is the synthetic image transmissivity.Meticulous skeleton view module obtains based on such thought, RGB min figure is actually the profile information of meticulous skeleton view, and the rough skeleton view that RGB_win_min forms is the approximate monochrome information of meticulous skeleton view, both generate meticulous skeleton view after merging by the guidance type image filter, and then obtain image transmissivity t.
The formula of guidance type image filter is as follows:
Among RGB_min figure and the RGB_win_min figure, the covariance of current calculation level place piece is:
cov_Ip＝mean_Ip-mean_I.*mean_p；
(wherein mean_I is the brightness of image mean value of current some place piece among the RGB_min figure;
Mean_p is the brightness of image mean value of current some place piece among the RGB_win_min figure;
Mean_Ip is the mean value of all corresponding point products of current some place piece among current some place piece and the RGB_win_min figure among the RGB_min figure; )
Among the RGB_min figure, the piece variance at current calculation level place is:
var_I＝mean_II-mean_I.*mean_I；
(wherein mean_I is the brightness of image mean value of current some place piece among the RGB_min figure;
Mean_II is the mean value of the brightness of image square of current some place piece among the RGB_min figure);
Likeness coefficient is:
a＝cov_Ip./(var_I+eps)；
(wherein eps prevents that for adjusting item var_I from being 0, and eps gets 0.01)
The otherness coefficient is:
b＝mean_p-a.*mean_I；
Meticulous skeleton view is:
q＝mean_a.*I+mean_b；
(wherein mean_a is the mean value of current some place piece of a, and wherein mean_b is the mean value of current some place piece of b)
The image transmissivity is:
t＝1-q/A；
(wherein A is atmosphere light)
As shown in Figure 6, the effect of this module is to generate palinspastic map.Had that " " A, degraded image I, image transmissivity t are according to outdoor objects imaging model I=J*t+A* (1-t) for atmosphere light; Palinspastic map J=(I-A* (1-t))/t.Image is comprised of three physical channels of RGB, so there are three identical module parallel synchronous to calculate.
As shown in Figure 7, the effect of this module is the brightness average of calculating input image, and determines that according to this value mode of operation is night vision pattern or day mode, if the brightness average of input picture is higher than a threshold value, be day mode, according to flow processing as mentioned above; Otherwise be the night vision pattern, to the image negate, then pass through after the above-mentioned flow process at the input end of image, again the image negate, can obtain like this low-light (level) hypograph brightness and contrast and promote.
Claims (6)
1. based on the weak signal extraction method, a kind of method that the video image of degenerating is restored: by the extraction to weak signal in the video image three primary colors, form an initial perspective view; By an efficient guidance type image filter, the gray-scale map modulation initial perspective view with raw video image forms the final skeleton view with edge details, and then obtains the image transmissivity again; Simultaneously, by being added up, the initial perspective view overall situation obtains a parameter that is called " atmosphere light "; Use at last image transmissivity, " atmosphere light " to carry out the anti-aliasing calculation process with former figure, restore the palinspastic map that visually details is very clear, color is very bright-coloured; If the brightness average of input picture is higher than a threshold value, according to flow processing as mentioned above; Otherwise at the input end of image, to the image negate, then pass through after the above-mentioned flow process, again the image negate; Can obtain like this low-light (level) hypograph brightness and contrast promotes.
2. weak signal extraction method according to claim 1 method that the video image of degenerating is restored, its feature is: by the extraction to weak signal in the video image three primary colors, form an initial perspective view.
3. weak signal extraction method according to claim 1 method that the video image of degenerating is restored, its feature is: by an efficient guidance type image filter, gray-scale map modulation initial perspective view with raw video image, formation has the final skeleton view of edge details, and then obtains the image transmissivity.
4. weak signal extraction method according to claim 1 method that the video image of degenerating is restored, its feature is: obtain a parameter that is called " atmosphere light " by the initial perspective view overall situation is added up.
5. weak signal extraction method according to claim 1 method that the video image of degenerating is restored, its feature is: carry out the anti-aliasing calculation process with meticulous skeleton view, " atmosphere light " with former figure, restore the palinspastic map that visually details is very clear, color is very bright-coloured.
6. weak signal extraction method according to claim 1 method that the video image of degenerating is restored, its feature is: if the brightness average of input picture is lower than a threshold value, be at input end and the output terminal of image, to the image negate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100940319A CN103366345A (en) | 2012-03-31 | 2012-03-31 | Degraded video image restoration technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012100940319A CN103366345A (en) | 2012-03-31 | 2012-03-31 | Degraded video image restoration technology |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN103366345A true CN103366345A (en) | 2013-10-23 |
Family
ID=49367620
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012100940319A Pending CN103366345A (en) | 2012-03-31 | 2012-03-31 | Degraded video image restoration technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103366345A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107644409A (en) * | 2017-09-28 | 2018-01-30 | 深圳Tcl新技术有限公司 | Image enchancing method, display device and computer-readable recording medium |
| CN112071037A (en) * | 2019-06-11 | 2020-12-11 | 陈军 | Site indicator lamp driving method |
| CN113379640A (en) * | 2021-06-25 | 2021-09-10 | 哈尔滨工业大学 | Multistage filtering image denoising method fusing edge information |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101950416A (en) * | 2010-09-15 | 2011-01-19 | 北京理工大学 | Bidirectional filtration-based real-time image de-hazing and enhancing method |
| US20110188775A1 (en) * | 2010-02-01 | 2011-08-04 | Microsoft Corporation | Single Image Haze Removal Using Dark Channel Priors |
| CN102254313A (en) * | 2011-07-14 | 2011-11-23 | 浙江大学 | Image defogging method based on restoration and fusion of images on foggy days |
-
2012
- 2012-03-31 CN CN2012100940319A patent/CN103366345A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110188775A1 (en) * | 2010-02-01 | 2011-08-04 | Microsoft Corporation | Single Image Haze Removal Using Dark Channel Priors |
| CN101950416A (en) * | 2010-09-15 | 2011-01-19 | 北京理工大学 | Bidirectional filtration-based real-time image de-hazing and enhancing method |
| CN102254313A (en) * | 2011-07-14 | 2011-11-23 | 浙江大学 | Image defogging method based on restoration and fusion of images on foggy days |
Non-Patent Citations (1)
| Title |
|---|
| 杨靖宇: "利用暗原色先验知识实现航空影像快速去雾", 《武汉大学学报·信息科学版》, vol. 35, no. 11, 30 November 2010 (2010-11-30), pages 1292 - 1295 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107644409A (en) * | 2017-09-28 | 2018-01-30 | 深圳Tcl新技术有限公司 | Image enchancing method, display device and computer-readable recording medium |
| CN112071037A (en) * | 2019-06-11 | 2020-12-11 | 陈军 | Site indicator lamp driving method |
| CN112071037B (en) * | 2019-06-11 | 2022-07-01 | 虹球环艺科技集团有限公司 | Method for driving field indicator lamp |
| CN113379640A (en) * | 2021-06-25 | 2021-09-10 | 哈尔滨工业大学 | Multistage filtering image denoising method fusing edge information |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xu et al. | Removing rain and snow in a single image using guided filter | |
| CN106251301B (en) | A kind of single image to the fog method based on dark primary priori | |
| Xu et al. | An improved guidance image based method to remove rain and snow in a single image | |
| CN107680056B (en) | Image processing method and device | |
| US8515196B1 (en) | Systems and methods for processing infrared images | |
| CN107292830B (en) | Low-illumination image enhancement and evaluation method | |
| CN106897981A (en) | A kind of enhancement method of low-illumination image based on guiding filtering | |
| CN103020920A (en) | Method for enhancing low-illumination images | |
| CN102750674A (en) | Video image defogging method based on self-adapting allowance | |
| CN108154492B (en) | A kind of image based on non-local mean filtering goes haze method | |
| CN104349069A (en) | Method for shooting high dynamic range film | |
| CN104021532A (en) | Image detail enhancement method for infrared image | |
| CN102063704B (en) | Airborne vision enhancement method and device | |
| CN107360344B (en) | Rapid defogging method for monitoring video | |
| GB2514895A (en) | A method and an apparatus for reducing image noise | |
| CN104537634A (en) | Method and system for removing raindrop influences in dynamic image | |
| CN104618700A (en) | Enhanced display method for color high dynamic range image | |
| CN113822826B (en) | Low-illumination image brightness enhancement method | |
| CN104506755A (en) | Method for real-time automatic defogging of high-definition videos based on FPGA | |
| CN101626454A (en) | Method for intensifying video visibility | |
| CN112435184B (en) | Image recognition method for haze days based on Retinex and quaternion | |
| CN106846258A (en) | A kind of single image to the fog method based on weighted least squares filtering | |
| CN103366345A (en) | Degraded video image restoration technology | |
| Zhang et al. | A fast video image defogging algorithm based on dark channel prior | |
| CN104463812A (en) | Method for repairing video image disturbed by raindrops in shooting process |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20131023 |