CN105005966A - Haze-removing algorithm of single image based on physical properties of yellow haze - Google Patents
Haze-removing algorithm of single image based on physical properties of yellow haze Download PDFInfo
- Publication number
- CN105005966A CN105005966A CN201510023351.9A CN201510023351A CN105005966A CN 105005966 A CN105005966 A CN 105005966A CN 201510023351 A CN201510023351 A CN 201510023351A CN 105005966 A CN105005966 A CN 105005966A
- Authority
- CN
- China
- Prior art keywords
- image
- channel
- pixel
- haze
- numerical value
- 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.)
- Granted
Links
Landscapes
- Image Processing (AREA)
- Facsimile Image Signal Circuits (AREA)
Abstract
The invention discloses a haze-removing algorithm of a single image based on physical properties of yellow haze, and belongs to the field of image processing. The method comprises: obtaining a to-be-processed image; determining a sky area of the to-be-processed image; determining atmospheric light spots in the sky area; determining a specific value of a color channel in the to-be-processed image; determining a transmission diagram of the to-be-processed image according to the specific value of the color channel in the to-be-processed image; adjusting saturability of the transmission diagram according to the physical properties of yellow haze; and generating an adjusted image. Compared with a method of uniformly processing fog and haze in the prior art, the method can adjust color of the image to be quite similar to the real color, so that original brightness and saturability of the image are recovered.
Description
Technical field
The present invention relates to image processing field, particularly a kind of single image based on yellow haze physical characteristics removes haze algorithm.
Background technology
Nowadays air pollution is day by day serious, especially haze weather, and after entering the winter, almost every day all can occur, and has scattering process, make image detail fuzzy at the picture of outdoor shooting due to haze particle to light, and overall image quality declines.
In the prior art, the image processing method of main flow obtains the air light value in picture, the distribution plan of propagation in atmosphere transmissivity in this image is determined according to air light value, and then according to propagation in atmosphere transmissivity determination mist elimination expression formula, to process image according to above-mentioned mist elimination haze expression formula, thus reach removal haze sky to the impact of image.
But inventor finds to there is following problem in prior art:
Choose by the highest brightness value of brightness in naked eyes determination image to air light value in prior art, there is no the accurate meaning considering brightness in atmospherical scattering model, and weather conditions not different with haze these two kinds to mist in the process of image procossing is carried out differentiation and is treated, but carry out unified process, because mist is different with the concrete origin cause of formation of haze, the result of unified processing mode is adopted to be to reach the effect of correcting aberration, the original brightness and contrast of Recovery image like this.
Summary of the invention
In order to solve the problem of prior art, the invention provides a kind of single image based on yellow haze physical characteristics and remove haze algorithm, described single image goes haze algorithm to comprise:
Obtain pending image, determine the sky areas in described pending image;
In described sky areas, determine air luminous point, determine the ratio of color channel in described pending image;
According to the ratio of color channel in described pending passage, determine the transmission diagram of described pending image;
In conjunction with the physical characteristics of yellow haze, regulate the saturation degree of described transmission diagram, generate the image after regulating.
Optionally, the pending image of described acquisition, determine the sky areas in described pending image, comprising:
Described pending image is divided into multiple region, extracts the first area in described pending image, Color edge detection is carried out to described first area, obtains edge image;
Binary conversion treatment is carried out to described edge image, obtains the near edge image after processing;
Detect in described near edge image, if meet first pre-conditioned, then judge that described first area that described near edge figure is corresponding is as sky areas.
Optionally, describedly in described sky areas, determine air luminous point, determine the ratio of color channel in described pending image, comprising:
The highest point of brightness is chosen as air luminous point in sky areas described at least one;
Determine the first number ratios of red channel, green channel, blue channel in described air luminous point, using described first number ratios as color channel ratio in described pending image.
Optionally, the described ratio according to color channel in described pending passage, determine the transmission diagram of described pending image, comprising:
According to the ratio of color channel in described pending image, by the first adjustment formula, the numerical value of each pixel in red channel, green channel and blue channel in described pending image is adjusted, numerical value after being adjusted, described first adjustment formula is specially:
Wherein, R' is the numerical value of red channel after the adjustment of each pixel, G' is the numerical value of green channel after the adjustment of each pixel, B' is the numerical value of blue channel after the adjustment of each pixel, R is the numerical value of red channel before the adjustment of each pixel, G is the numerical value of green channel before the adjustment of each pixel, and B is the numerical value of blue channel before the adjustment of each pixel, R
wfor the numerical value of air luminous point red channel, G
wfor the numerical value of air luminous point green channel, B
wfor the numerical value of atmosphere light point blue channel, T is the default passage value upper limit;
Based on red, green, blue scattered power and wavelength relation, in conjunction with human eye to described redness, green, blue sensitivity, determine to be converted to HSI (Hue-Saturation-Intensity at color space, colourity-saturation degree-intensity) time intensity-conversion formula, obtain the intensity level of each pixel in described pending image according to described intensity-conversion formula, described intensity-conversion formula is specially:
I=0.4520*R'+0.5121*G'+0.0359*B',
Wherein, I is the intensity level in HSI color space;
After being converted to described HSI color space, choose the minimum value of each pixel numerical value in described red channel, green channel, blue channel in described pending image, according to described minimum value composition gray-scale map, bilateral filtering is carried out to described gray-scale map, forms the transmission diagram of described pending image.
Optionally, the physical characteristics of the yellow haze of described combination, regulates the saturation degree of described transmission diagram, generates the image after regulating, comprising:
In conjunction with the second adjustment formula, regulate the saturation degree of pixel each in described transmission diagram, generate the image after regulating, described second adjustment formula is specially:
Wherein, S is the initial saturation of each pixel in described pending image, and S' is the saturation degree of each pixel after process.
The beneficial effect that technical scheme provided by the invention is brought is:
By determining the sky areas in pending image, and then choose the standard point of the most bright spot in sky areas as white balance, then the color adjustment of full image is carried out according to the color channel ratio of this standard point, and carry out the conversion of color space, and according to the characteristic of yellow haze, the saturation degree after conversion is adjusted, avoid the generation can not implementing adjustment in prior art for the characteristic of haze to image color, reduce image aberration, as far as possible the original brightness and contrast of Recovery image.
Accompanying drawing explanation
In order to be illustrated more clearly in technical scheme of the present invention, below the accompanying drawing used required in describing embodiment is briefly described, apparently, accompanying drawing in the following describes is only some embodiments of the present invention, for those of ordinary skill in the art, under the prerequisite not paying creative work, other accompanying drawing can also be obtained according to these accompanying drawings.
Fig. 1 is the schematic flow sheet that a kind of single image based on yellow haze physical characteristics provided by the invention removes haze algorithm;
Fig. 2 is the contrast schematic diagram that a kind of single image based on yellow haze physical characteristics provided by the invention removes the filtering mode in haze algorithm;
Fig. 3 is the position view of the original image of sample 1, near edge image and air luminous point;
Fig. 4 is the position view of the original image of sample 2, near edge image and air luminous point;
Fig. 5 is the position view of the original image of sample 3, near edge image and air luminous point;
Contrast histogram before and after the original image of Fig. 6 sample 4, color balance simulation result and emulation;
Contrast histogram before and after the original image of Fig. 7 sample 5, color balance simulation result and emulation;
Fig. 8 is the original image of sample 6 and the image after going haze to emulate;
Fig. 9 is the original image of sample 7 and the image after going haze to emulate.
Embodiment
For making structure of the present invention and advantage clearly, below in conjunction with accompanying drawing, structure of the present invention is further described.
Embodiment one
The present embodiment provides a kind of single image based on yellow haze physical characteristics to remove haze algorithm, and as shown in Figure 1, described single image goes haze algorithm to comprise:
101, obtain pending image, determine the sky areas in described pending image.
102, in described sky areas, determine air luminous point, determine the ratio of color channel in described pending image.
103, according to the ratio of color channel in described pending passage, the transmission diagram of described pending image is determined.
104, in conjunction with the physical characteristics of yellow haze, regulate the saturation degree of described transmission diagram, generate the image after regulating.
In force, first in pending image, sky areas is determined, after determining air luminous point in region on high, the ratio of color channel in pending image is determined according to fixed air luminous point, and then the transmission diagram of pending image is determined according to the ratio of color channel, finally according to the physical characteristics of yellow haze, the saturation degree of described transmission diagram is regulated, generate the image after regulating.Go in haze process in entirety, owing to choosing the standard that in sky areas, air luminous point adjusts as the color channel in pending image, like this can the original color of Recovery image as far as possible, adding that the saturation degree of physical characteristics to image for yellow haze regulates, like this can by mist and the unified mode processed of haze relatively and in prior art, can by image adjustment to closer to real color, thus the original brightness of Recovery image and saturation degree.
Optionally, the pending image of described acquisition, determine the sky areas in described pending image, comprising:
Step one, is divided into multiple region by described pending image, extracts the first area in described pending image, carries out Color edge detection, obtain edge image to described first area.
In force, according to investigation statistics analysis, the picture of major part under yellow haze weather is outdoor shooting by day, and in environment outdoor by day, namely the main source of illumination is direct sunlight and the diffuse reflection daylight, the region that corresponding intensity of illumination is the highest is exactly sky areas, other part blocking more or less due to building or other objects, corresponding intensity of illumination all can not higher than sky areas, under these circumstances, even if there is the interference of yellow haze, the loss of the brightness value of sky areas also can not be too high, therefore in pending image, sky areas is determined, just become a very important problem.
In above-mentioned steps one, in advance Region dividing is carried out to pending image, obtain multiple region, the image at every turn choosing region corresponding carries out Color edge detection, concrete Color edge detection method can use the ColorGrad algorithm calculating color gradient, carry out rim detection for the coloured image in rgb color space, obtain edge image, this edge image is gray level image.
Step 2, carries out binary conversion treatment to described edge image, obtains the near edge image after processing.
In force, because edge image is gray level image, comprise multiple excessive GTG look, for the ease of post-processed, here binaryzation principle is introduced, binary conversion treatment is carried out to the edge image obtained in step one, be exactly briefly set up to distinguish a threshold value, by the pixel assignment 1 higher than this differentiation threshold value in gray-scale value corresponding for pixel each in edge image, be white, the pixel assignment 0 of this differentiation threshold value will be equal to or less than in gray-scale value corresponding for each pixel, be black, what obtain like this is the outline map only having black and white two look, the edge image a kind of relative to step is more succinctly meticulous, be referred to as near edge image.
Step 3, detects in described near edge image, if meet first pre-conditioned, then judges that described first area that described near edge figure is corresponding is as sky areas.
In force, detect in conjunction with the first pre-conditioned near edge image to having generated, here first is pre-conditionedly specially:
(1) assignment that pixel is corresponding is 0;
(2) luminance component that pixel is corresponding is greater than predetermined threshold value I
t;
(3) the saturation degree component that pixel is corresponding is less than predetermined threshold value S
t;
Wherein, predetermined threshold value I
t=0.65*I
max+ 0.35*I
min,
I
maxand I
minbe respectively maximal value and the minimum value of region P luminance component, S
medfor the intermediate value of region P saturation degree component.
For the ease of understanding, be described in conjunction with concrete data here.Such as, choose in pending image and be positioned at (50,380) pixel processes, the numerical value of this pixel RGB namely respectively in red channel, green channel, blue channel is (0.8902,0.8471,0.7607), the predetermined threshold value formula in pre-conditioned according to first, predetermined threshold value I
maxbe 0.7843, I
minbe 0.2157, S
medbe 0.2298, in conjunction with the I that above-mentioned computing formula obtains
t=0.5853, St=0.2298, for the I=0.7608>I of (50,380) pixel
ts=0.0863<St, meets the requirements, and this pixel therefore in first area is by the first pre-conditioned detection, if pixels all in this first area all passes through the first pre-conditioned detection by above-mentioned steps, then this first area is chosen to be sky areas.
It is worth mentioning that, due to when normally taking pictures, sky areas is be in the upper board part of pending image substantially, so a kind of first area chosen of step is preferably the first half of pending image in most cases, if do not meet the first pre-conditioned sky areas in the place of choosing through above-mentioned steps one to step 3 in the first region, can also according to actual user demand, consider the decision process the need of choosing other parts and again carrying out as shown in step one to step 3, till selecting and meeting the first pre-conditioned sky areas.By above-mentioned steps once choosing sky areas from pending image to step 3, and proceed process for step afterwards.
Optionally, describedly in described sky areas, determine air luminous point, determine the ratio of color channel in described pending image, comprising:
The highest point of brightness is chosen as air luminous point in sky areas described at least one;
Determine the first number ratios of red channel, green channel, blue channel in described air luminous point, using the ratio of described first number ratios as color channel in described pending image.
In force, the sky areas may determined in step 3 has multiple, in at least one sky areas determined, choose a highest point of brightness as air luminous point, then the numerical value of this air luminous point in pending image respectively in red channel, green channel and blue channel three passages is obtained, and be as the criterion with the numerical value of green channel, determine red channel, the blue channel number ratios relative to green channel respectively.
Here why use self white balance of air luminous point as the standard of pending image adjustment white balance, because air luminous point has the highest brightness value in pending image, relative to other pixels, the effect affected by yellow haze is minimum, goes to regulate pending image can have best regulating effect with the white balance of this air luminous point.
Optionally, the described ratio according to color channel in described pending passage, determine the transmission diagram of described pending image, comprising:
According to the ratio of color channel in described pending image, by the first adjustment formula, the numerical value of each pixel in red channel, green channel and blue channel in described pending image is adjusted, numerical value after being adjusted, described first adjustment formula is specially:
Wherein, R' is the numerical value of red channel after the adjustment of each pixel, G' is the numerical value of green channel after the adjustment of each pixel, B' is the numerical value of blue channel after the adjustment of each pixel, R is the numerical value of red channel before the adjustment of each pixel, G is the numerical value of green channel before the adjustment of each pixel, and B is the numerical value of blue channel before the adjustment of each pixel, R
wfor the numerical value of air luminous point red channel, G
wfor the numerical value of air luminous point green channel, B
wfor the numerical value of atmosphere light point blue channel, T is the default passage value upper limit;
Based on red, green, blue scattered power and wavelength relation, in conjunction with human eye to described redness, green, blue sensitivity, determine to be converted to HSI (Hue-Saturation-Intensity at color space, colourity-saturation degree-intensity) time intensity-conversion formula, obtain the intensity level of each pixel in described pending image according to described intensity-conversion formula, described intensity-conversion formula is specially:
I=0.4520*R'+0.5121*G'+0.0359*B',
Wherein, I is the intensity level in HSI color space;
After being converted to described HSI color space, choose the minimum value of each pixel numerical value in described red channel, green channel, blue channel in described pending image, according to described minimum value composition gray-scale map, bilateral filtering is carried out to described gray-scale map, forms the transmission diagram of described pending image.
In force, the numerical value of the air luminous point determined before respectively in red channel, green channel, blue channel is substituted into respectively the R in the first adjustment formula
w, G
w, B
win, the numerical value of other pixels in red channel, green channel, blue channel in pending image is substituted in R, G, B respectively, in conjunction with default passage value upper limit T, just can determine the numerical value in the red channel after the adjustment of each pixel, green channel, blue channel according to the first adjustment formula, namely complete the step of according to air luminous point white balance, the pending image of view picture being carried out to blank level adjustment.
Owing to being mixed with a large amount of molecule such as water vapor, dust in an atmosphere, whole atmospheric environment is made to form the colloid of " gasoloid ", thus can produce the impact of scattering to the light through air, and for without light corresponding to wavelength, the degree of scattering is also not quite similar.In fact based on the difference having each wave band coloured light caused due to Rayleigh scattering under haze weather to contribute transfer rate, the relation of scattered power and wavelength under Rayleigh scattering condition is considered, the computing formula in conjunction with following transfer rate:
t(x)=e
-βd(x),
The ratio that can obtain transfer rate is about: 0.6498:0.3679:0.1514.
Reference human eye is to the difference in perception of three primary colours light simultaneously, with regard to RGB three primary colours, the brightness of human eye perceives green glow is the strongest, ruddiness takes second place (being about the half of green glow), blue light the most weak (being about 1/3rd of ruddiness), therefore, supposes that the vision light intensity of green glow is 1, the vision light intensity of ruddiness can only be 1/2 of its former intensity, and blue light can only be then 1/6 of its former intensity.
Further, treatment step before all processes based on rgb color space, oversaturated phenomenon is easily there is at this color space hypograph, therefore in order to realize the accurate reparation of image color, here the conversion carrying out color space is needed, namely be converted to HSI (Hue-Saturation-Intensity, colourity-saturation degree-intensity) from RGB, thus realize the accurate control of saturation degree.
Comprehensively above-mentioned three reasons, the intensity-conversion formula finally determined under HSI color space in conjunction with the rgb value after each pixel adjustment is:
I=0.4520*R'+0.5121*G'+0.0359*B',
Wherein, R' is the numerical value of red channel after the adjustment of each pixel, and G' is the numerical value of green channel after the adjustment of each pixel, and B' is the numerical value of blue channel after the adjustment of each pixel, in HSI color space, determined intensity (i.e. brightness) value of each pixel by above-mentioned intensity-conversion formula.
Further, after being converted to HSI color space, choose the minimum value of each pixel numerical value in described red channel, green channel, blue channel in pending image, according to described minimum value composition gray-scale map, bilateral filtering is carried out to described gray-scale map, forms the transmission diagram of described pending image.
Above-mentioned gray-scale map is exactly inherently a gray-scale map, the gray-scale value of each pixel is determined by the minimum value of pixel each in transmission diagram in three kinds of Color Channels, if namely the numerical value of a pixel in red channel, green channel, blue channel is respectively (0.8902,0.8471,0.7607) wherein minimum value 0.7607, is then got as gray-scale value corresponding to this pixel in gray-scale map.
It is worth mentioning that, here bilateral filtering is for the mini-value filtering conventional relative to classic method, by processing above-mentioned gray-scale map in spatial domain and codomain simultaneously, specifically as shown in Figure 2, for pixel A, mini-value filtering obtains centered by pixel A, around the pixel value of adjacent 8 pixels, after above-mentioned 8 pixel values are asked for minimum value, using the value of minimum value as pixel A; Bilateral filtering is then after by 8 pixel values adjacent around pixel A, and the weight a-h in conjunction with each value calculates, using the numerical value that the obtains value as pixel A.Bilateral filtering, relative to mini-value filtering, the value to pixel A can process from spatial domain and codomain two aspect, avoid the defect that mini-value filtering fully can not reflect surrounding pixel value.
Optionally, the physical characteristics of the yellow haze of described combination, regulates the saturation degree of described transmission diagram, generates the image after regulating, comprising:
In conjunction with the second adjustment formula, regulate the saturation degree of pixel each in described transmission diagram, generate the image after regulating, described second adjustment formula is specially:
Wherein, S is the initial saturation of each pixel in described pending image, and S' is the saturation degree of each pixel after process.
In force, the aerosol systems that is made up of the small water droplet be suspended in a large number in surface air or ice crystal of mist.Haze is because the particles such as the dust in air, sulfuric acid, nitric acid, organic hydrocarbon compounds make the phenomenon of air muddiness.As can be seen here, mist and haze are in essence and different, targetedly the image under yellow haze weather is processed in the present embodiment, therefore, in conjunction with the physical characteristics of yellow haze, need to regulate the saturation degree of the transmission diagram that previous step generates, specifically according to the second adjustment formula, the saturation degree S of pixel each in transmission diagram is carried out classified calculating according to whether being greater than 0.025, thus obtain the intensity value of each pixel after calculating, after the saturation degree of pixel each in transmission diagram is replaced with the intensity value after calculating, the image after adjusted.
It should be noted that, span is in practice 0 to 255, and the span of the numerical value in the present embodiment in red channel, green channel, blue channel is 0 to 1, is because the numerical value of 0 to 255 all have been done the process divided by 255, in order to avoid misreading, spy explains herein.
In order to show the process advantage of this method, concrete to realize effect as follows:
Emulation 1, to the emulation of atmosphere light estimation algorithm.
Emulating 1 simulated conditions is carry out under MATLAB R2008a software.
Estimation atmosphere light is carried out to test sample book 1-3 and carries out emulation experiment.
Fig. 3 is the position view of the original image of sample 1, near edge image and air luminous point;
Fig. 4 is the position view of the original image of sample 2, near edge image and air luminous point;
Fig. 5 is the position view of the original image of sample 3, near edge image and air luminous point.
In as can be seen from Fig. 3 to Fig. 5, the interference of white object can well be avoided for sample 1, avoid the mistake detecting leakage problem that traditional algorithm causes owing to ignoring atmosphere light physical significance; For sample 2, due to the color information paying close attention to image, thus avoid the interference because interpolation produces in image depth sudden change place preferably, correctly brightness abnormity point is got rid of on high outside part; For sample 3, owing to detecting in the brightness of image and saturation degree component, thus avoid the interference of artificial light sources simultaneously.
Emulation 2, to the emulation of colour-balance algorithm.
Emulating 2 simulated conditions is carry out under MATLAB R2008a software.
Color balance emulation experiment is carried out to test sample book 4-5.
Contrast histogram before and after the original image of Fig. 6 sample 4, color balance simulation result and emulation;
Contrast histogram before and after the original image of Fig. 7 sample 5, color balance simulation result and emulation.
As can be seen from Fig. 6 and 7, the present invention proposes colour-balance algorithm and has good adaptivity.For the image such as Fig. 6 that there is not color distortion, as can be seen from Fig. 6, in original image and emulating image histogram, curve overlaps substantially, and this algorithm can't be made an amendment, and can well keep image color character originally; For color distortion comparatively significantly such as Fig. 7, from Fig. 7 in original image and emulating image histogram, the curve of original image original image and emulating image refined image has many places not overlap can to find out, this algorithm has certain change to information such as the brightness of image on a balanced basis.
Emulation 3, to the emulation of final effect of removing haze.
Emulating 3 simulated conditions is carry out under MATLAB R2008a software.
Emulation experiment is carried out to test sample book.Fig. 8 is the original image of sample 6 and the image after going haze to emulate, and Fig. 9 is the original image of sample 7 and the image after going haze to emulate.
As can be seen from Fig. 8,9, the algorithm in the present invention achieves and good goes haze effect.The color distortion brought by dense haze can well be eliminated for this algorithm of sample, and there will not be supersaturation or halo (the abnormal halation depth of field sudden change place artificially produces) phenomenon; The detailed information of local can be retained well in the place that the depth of field is darker for this algorithm of sample.
A kind of single image based on yellow haze physical characteristics proposed in the present embodiment removes haze algorithm, by determining air luminous point in pending image, the ratio of color channel in pending image is determined according to fixed air luminous point, and then the transmission diagram of pending image is determined according to the ratio of color channel, finally according to the physical characteristics of yellow haze, the saturation degree of described transmission diagram is regulated, generates the image after regulating.Can the original color of Recovery image as far as possible, and regulate for the saturation degree of physical characteristics to image of yellow haze, like this can by mist and the unified mode processed of haze relatively and in prior art, can by image adjustment to closer to real color, thus the original brightness of Recovery image and saturation degree.
The foregoing is only embodiments of the invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (5)
1. the single image based on yellow haze physical characteristics removes a haze algorithm, it is characterized in that, described single image goes haze algorithm to comprise:
Obtain pending image, determine the sky areas in described pending image;
In described sky areas, determine air luminous point, determine the ratio of color channel in described pending image;
According to the ratio of color channel in described pending passage, determine the transmission diagram of described pending image;
In conjunction with the physical characteristics of yellow haze, regulate the saturation degree of described transmission diagram, generate the image after regulating.
2. the single image based on yellow haze physical characteristics according to claim 1 removes haze algorithm, it is characterized in that, the pending image of described acquisition is determined the sky areas in described pending image, being comprised:
Described pending image is divided into multiple region, extracts the first area in described pending image, Color edge detection is carried out to described first area, obtains edge image;
Binary conversion treatment is carried out to described edge image, obtains the near edge image after processing;
Detect in described near edge image, if meet first pre-conditioned, then judge that described first area that described near edge figure is corresponding is as sky areas.
3. the single image based on yellow haze physical characteristics according to claim 1 removes haze algorithm, it is characterized in that, describedly in described sky areas, determines air luminous point, determines the ratio of color channel in described pending image, comprising:
The highest point of brightness is chosen as air luminous point in sky areas described at least one;
Determine the first number ratios of red channel, green channel, blue channel in described air luminous point, using described first number ratios as color channel ratio in described pending image.
4. the single image based on yellow haze physical characteristics according to claim 1 removes haze algorithm, it is characterized in that, the described ratio according to color channel in described pending passage, determines the transmission diagram of described pending image, comprising:
According to the ratio of color channel in described pending image, by the first adjustment formula, the numerical value of each pixel in red channel, green channel and blue channel in described pending image is adjusted, numerical value after being adjusted, described first adjustment formula is specially:
Wherein, R' is the numerical value of red channel after the adjustment of each pixel, G' is the numerical value of green channel after the adjustment of each pixel, B' is the numerical value of blue channel after the adjustment of each pixel, R is the numerical value of red channel before the adjustment of each pixel, G is the numerical value of green channel before the adjustment of each pixel, and B is the numerical value of blue channel before the adjustment of each pixel, R
wfor the numerical value of air luminous point red channel, G
wfor the numerical value of air luminous point green channel, B
wfor the numerical value of atmosphere light point blue channel, T is the default passage value upper limit;
Based on red, green, blue scattered power and wavelength relation, in conjunction with human eye to described redness, green, blue sensitivity, determine to be converted to HSI (Hue-Saturation-Intensity at color space, colourity-saturation degree-intensity) time intensity-conversion formula, obtain the intensity level of each pixel in described pending image according to described intensity-conversion formula, described intensity-conversion formula is specially:
I=0.4520*R'+0.5121*G'+0.0359*B',
Wherein, I is the intensity level in HSI color space;
After being converted to described HSI color space, choose the minimum value of each pixel numerical value in described red channel, green channel, blue channel in described pending image, according to described minimum value composition gray-scale map, bilateral filtering is carried out to described gray-scale map, forms the transmission diagram of described pending image.
5. the single image based on yellow haze physical characteristics according to claim 1 removes haze algorithm, it is characterized in that, the physical characteristics of the yellow haze of described combination, regulates the saturation degree of described transmission diagram, generate the image after regulating, comprising:
In conjunction with the second adjustment formula, regulate the saturation degree of pixel each in described transmission diagram, generate the image after regulating, described second adjustment formula is specially:
Wherein, S is the initial saturation of each pixel in described pending image, and S' is the saturation degree of each pixel after process.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510023351.9A CN105005966B (en) | 2015-01-16 | 2015-01-16 | A kind of single image based on yellow haze physical characteristic goes haze method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510023351.9A CN105005966B (en) | 2015-01-16 | 2015-01-16 | A kind of single image based on yellow haze physical characteristic goes haze method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105005966A true CN105005966A (en) | 2015-10-28 |
CN105005966B CN105005966B (en) | 2017-07-28 |
Family
ID=54378625
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510023351.9A Active CN105005966B (en) | 2015-01-16 | 2015-01-16 | A kind of single image based on yellow haze physical characteristic goes haze method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105005966B (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106954022A (en) * | 2017-03-08 | 2017-07-14 | 广东欧珀移动通信有限公司 | Image processing method, device and terminal |
CN106998445A (en) * | 2017-03-28 | 2017-08-01 | 深圳市科美集成电路有限公司 | Haze penetrates camera circuit and system |
CN109361907A (en) * | 2018-11-16 | 2019-02-19 | 深圳市趣创科技有限公司 | Method, apparatus, terminal and the storage medium of camera shooting night scene sky color |
CN109919859A (en) * | 2019-01-25 | 2019-06-21 | 暨南大学 | A kind of Outdoor Scene image defogging Enhancement Method calculates equipment and its storage medium |
CN112330559A (en) * | 2020-11-05 | 2021-02-05 | 山东交通学院 | Early warning method for image information recovery and lane keeping of severe foggy roads |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7409082B2 (en) * | 1999-08-18 | 2008-08-05 | Fujifilm Corporation | Method, apparatus, and recording medium for processing image data to obtain color-balance adjusted image data based on white-balance adjusted image data |
CN101908210A (en) * | 2010-08-13 | 2010-12-08 | 北京工业大学 | Method and system for color image defogging treatment |
CN103150708A (en) * | 2013-01-18 | 2013-06-12 | 上海交通大学 | Image quick defogging optimized method based on black channel |
CN103914813A (en) * | 2014-04-10 | 2014-07-09 | 西安电子科技大学 | Colorful haze image defogging and illumination compensation restoration method |
-
2015
- 2015-01-16 CN CN201510023351.9A patent/CN105005966B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7409082B2 (en) * | 1999-08-18 | 2008-08-05 | Fujifilm Corporation | Method, apparatus, and recording medium for processing image data to obtain color-balance adjusted image data based on white-balance adjusted image data |
CN101908210A (en) * | 2010-08-13 | 2010-12-08 | 北京工业大学 | Method and system for color image defogging treatment |
CN103150708A (en) * | 2013-01-18 | 2013-06-12 | 上海交通大学 | Image quick defogging optimized method based on black channel |
CN103914813A (en) * | 2014-04-10 | 2014-07-09 | 西安电子科技大学 | Colorful haze image defogging and illumination compensation restoration method |
Non-Patent Citations (4)
Title |
---|
FAN GUO ET AL.: "Image Dehazing Based on Haziness Analysis", 《INTERNATIONAL JOURNAL OF AUTOMATION AND COMPUTING》 * |
JING YU ET AL.: "Physics-based Fast Single Image Fog Removal", 《ICS 2010 PROCEEDINGS》 * |
SYEDZ.MASOOD ET AL.: "Automatic Correction of Saturated Regions in Photographs using Cross-channel Correlation", 《PACIFIC GRAPHICS 2009》 * |
胡学友: "雾天降质图像的增强复原算法研究", 《中国博士学位论文全文数据库 信息科技辑》 * |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106954022A (en) * | 2017-03-08 | 2017-07-14 | 广东欧珀移动通信有限公司 | Image processing method, device and terminal |
CN106954022B (en) * | 2017-03-08 | 2019-10-25 | Oppo广东移动通信有限公司 | Image processing method, device and terminal |
CN106998445A (en) * | 2017-03-28 | 2017-08-01 | 深圳市科美集成电路有限公司 | Haze penetrates camera circuit and system |
CN109361907A (en) * | 2018-11-16 | 2019-02-19 | 深圳市趣创科技有限公司 | Method, apparatus, terminal and the storage medium of camera shooting night scene sky color |
CN109919859A (en) * | 2019-01-25 | 2019-06-21 | 暨南大学 | A kind of Outdoor Scene image defogging Enhancement Method calculates equipment and its storage medium |
CN109919859B (en) * | 2019-01-25 | 2021-09-07 | 暨南大学 | Outdoor scene image defogging enhancement method, computing device and storage medium thereof |
CN112330559A (en) * | 2020-11-05 | 2021-02-05 | 山东交通学院 | Early warning method for image information recovery and lane keeping of severe foggy roads |
CN112330559B (en) * | 2020-11-05 | 2022-03-04 | 山东交通学院 | Early warning method for image information recovery and lane keeping of severe foggy roads |
Also Published As
Publication number | Publication date |
---|---|
CN105005966B (en) | 2017-07-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105005966A (en) | Haze-removing algorithm of single image based on physical properties of yellow haze | |
CN106056559A (en) | Dark-channel-prior-method-based non-uniform-light-field underwater target detection image enhancement method | |
CN106485681B (en) | Underwater color image restoration method based on color correction and red channel prior | |
CN104063853B (en) | Method for improving traffic video image definition based on dark channel technology | |
CN105654437A (en) | Enhancement method for low-illumination image | |
CN104766284B (en) | The restored method of underwater colored blur image | |
CN106157270A (en) | A kind of single image rapid defogging method and system | |
CN104463816A (en) | Image processing method and device | |
CN104867121B (en) | Image Quick demisting method based on dark primary priori and Retinex theories | |
CN104008528B (en) | Nonuniform illumination Underwater Target Detection image enchancing method based on Threshold segmentation | |
CN107316278A (en) | A kind of underwater picture clearness processing method | |
CN103793885B (en) | Compartmentalization image recovery method under inhomogeneous illumination in strong scattering optical imagery environment | |
CN107507145A (en) | A kind of underwater picture Enhancement Method based on the stretching of different colours spatially adaptive histogram | |
CN105631823A (en) | Dark channel sky area defogging method based on threshold segmentation optimization | |
CN105513019A (en) | Method and apparatus for improving image quality | |
CN110322410A (en) | Underwater picture defogging and color cast correction based on the compensation of bright channel transmissivity | |
CN107025881A (en) | The bearing calibration of display unit and device | |
CN102831590A (en) | Defogging strengthening method for transmissivity/dark channel prior based on spectral characteristic | |
CN104346783A (en) | Processing method and processing device for defogging image | |
CN104331867B (en) | The method, device and mobile terminal of image defogging | |
CN108257101A (en) | A kind of underwater picture Enhancement Method based on optimal recovery parameter | |
CN104021527A (en) | Rain and snow removal method in image | |
CN103971344B (en) | The colour of skin color cast correction of a kind of certificate image and system | |
CN110070480A (en) | A kind of analogy method of underwater optics image | |
CN104050678A (en) | Underwater monitoring color image quality measurement method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |