CN103426151B - A kind of image defogging method and device - Google Patents
A kind of image defogging method and device Download PDFInfo
- Publication number
- CN103426151B CN103426151B CN201310349564.1A CN201310349564A CN103426151B CN 103426151 B CN103426151 B CN 103426151B CN 201310349564 A CN201310349564 A CN 201310349564A CN 103426151 B CN103426151 B CN 103426151B
- Authority
- CN
- China
- Prior art keywords
- pixel
- value
- dark
- image
- present image
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a kind of image defogging method and device.The method includes: calculate the dark value of each pixel of present image；Wherein, the minima of described dark value is taken as global context light value；In present image, the straight line at two pixel places that described dark value is minimum and maximum is depth of field datum line, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；Calculate absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process.Adopting the present invention, it is possible to having reacted the gradual change law that " haze " is deep or light, the deep or light situation with reference to each pixel surrounding pixel avoids dark primary prior information to lose efficacy, thus obtaining better image treatment effect, and operand is few, treatment effeciency is high.
Description
Technical field
The present invention relates to image processing techniques, particularly relate to a kind of image defogging method and device.
Background technology
At present, method mist figure carrying out sharpening process generally has two classes: based on the algorithm with non-model of model.Wherein, carry out inverse operation based on the algorithm of model improve the contrast of image by understanding the immanent cause of image degradation；The algorithm of non-model is not required for knowing the information of image degradation reason.Image is processed we term it image enhaucament by the method for non-model；Processing image by the method based on model we term it image recovers, this kind of algorithm is relatively reliable, because they make use of the Physical Mechanism of image degradation, its object is to improve the fidelity of degraded image.
In recent years, the research of above-mentioned two aspect all achieves bigger progress, is based particularly on the recovery problem of physical model, has attracted the attention of more and more researcher, does brief introduction with regard to its present Research individually below.
Image enhaucament is the basic means of image procossing, refers to by some information specifically needed in prominent piece image, meanwhile, weakens or remove the processing method of some unwanted information.Picture superposition is a traditional topic in image processing field, is always up again the research field comparatively enlivened simultaneously, but, existing technology amount of calculation is huge, and computing is complicated, it is thus achieved that image processing effect not good yet.
Summary of the invention
Based on this, it is necessary to for the problems referred to above, it is provided that a kind of image defogging method and device, it is possible to obtain image comparatively clearly rapidly by a small amount of computing.
A kind of image defogging method, including:
Calculate the dark value of each pixel of present image；Wherein, the minima of described dark value is taken as global context light value；
In present image, the straight line at two pixel places that described dark value is minimum and maximum is depth of field datum line, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
Calculate absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process.
Correspondingly, a kind of image demister, including:
Dark computing unit, for calculating the dark value of each pixel of present image；
Environment light acquiring unit, is used for the minima taking described dark value as global context light value；
Environment passage acquiring unit, for with the straight line at minimum and maximum two the pixel places of described dark value in present image for depth of field datum line, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
The mist elimination processing unit being respectively connected with described environment light acquiring unit, described environment passage acquiring unit, for calculating absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process.
There is advantages that
First the present invention calculates the dark value of each pixel on image, then depth of field datum line is demarcated, described depth of field datum line has reacted the gradual change law that " haze " is deep or light, deep or light situation with reference to each pixel surrounding pixel avoids dark primary prior information to lose efficacy, thus obtaining better image treatment effect.The operand of the present invention is few, and treatment effeciency is high, it is possible to be widely used in the picture under haze weather, rainy day picture or even underwater photography picture and the Computer Vision under any of the above-described environment, in order to improve the definition of image.
Accompanying drawing explanation
Fig. 1 is the flow chart of image defogging method of the present invention；
Fig. 2 is the embodiment flow chart of image defogging method of the present invention；
Fig. 3 is the schematic diagram of image demister of the present invention；
Fig. 4 is the embodiment schematic diagram of image defogging method of the present invention.
Detailed description of the invention
For making the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail.
Fig. 1 is the flow chart of image defogging method of the present invention, including:
S101: calculate the dark value of each pixel of present image；Wherein, the minima of described dark value is taken as global context light value；
S102: the straight line at two pixel places that described dark value is minimum and maximum is depth of field datum line in present image, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
S103: calculate absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process.
The present invention is in the theoretical basis of physical model, and the imaging mechanism under severe atmospheric condition has carried out deep analysis the Enhancement Method that a kind of Misty Image proposed processes.Compared with strengthening with traditional images, this method is built upon, on the physical process of greasy weather imaging, therefore having more specific aim, and treatment effect is also ideal.
Based on the image that the single image mist elimination technology of prior information is not affected by fog in a large number by collection, it was found that a set of dark primary statistical law that can recognise that fog concentration.Namely divide the image into multiple sub-block, each sub-block has the pixel that some brightness are very low.During these " stains " are generally stored in shadow of object, black object and have the object of bright-colored.According to this rule, only by the color of fog concentration local route repair image each several part, just need to can effectively achieve good mist elimination effect, but when the brightness of scene objects is similar to atmosphere light, dark primary prior information will lose efficacy.This is also a difficult problem for prior art, and its adverse consequences is to cause image local overexposure, or the image generation variable color after mist elimination.The present invention, by the depth of field datum line mist elimination demarcated, avoids the generation of dark primary prior information failure conditions to a certain extent.Its principle is that in the photograph under haze weather, the distance of the depth of field just becomes positive correlation with the concentration of mist, photograph is selected 2 points that haze is the denseest and the lightest, with one depth of field datum line of this two-point drawing, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, carry out mist elimination process according to the physical model preset afterwards.
Compared to, the complicated processing method that some target edges along figure are sketched the contours, the present invention just can obtain good mist elimination effect by little amount of calculation, and treatment effeciency is high, and speed is fast.The picture under haze weather, rainy day picture or even underwater photography picture and the Computer Vision under any of the above-described environment can be widely used in, in order to improve the definition of image.
Fig. 2 is the embodiment flow chart of image defogging method of the present invention, and compared to Figure 1, Fig. 2 is the schematic diagram of the preferred embodiment of the present invention.
S201: calculate the dark value of each pixel of present image；
S202: travel through each pixel according to preset order, according to comprise current detection pixel default localized mass in the described dark value of each pixel, choose the minimum dark value dark value as the pixel correction of described current detection；
S203: take the minima of described dark value as global context light value；
S204: the straight line at two pixel places that described dark value is minimum and maximum is depth of field datum line in present image, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
S205: calculate absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process.
Calculate the dark value of each pixel, travel through according to default order from top to down or from left to right, according to comprise current detection pixel default localized mass in the described dark value of each pixel, choose the minimum dark value dark value I as the pixel correction of described current detectiondark(x).Described localized mass can be the border circular areas using the pixel of current detection as center, and rectangular area, is square area especially；Can also be (be positioned at the first row first row, coordinate is [1,1]) headed by the pixel of current detection, the matrix dot of composition or nine grids.
Wherein in the middle of an embodiment, choose centered by the pixel of described current detection, preset rectangular area long, default wide, as described localized mass.The present embodiment, centered by the pixel of current detection, fully takes into account " haze " concentration around this pixel.
Specifically, with the dark value I of current pixeldarkX () takes centered by current pixel, radius is that R(R is defaulted as 6) window in minimum dark channel value in all pixels, it may be assumed that
Wherein, ICY shadow passage (DarkChannel) that () is pixel is based primarily upon the observation of the outdoor images to fine day (haze-free): for most of pixel of non-sky areas, has at least a Color Channel to have low-down brightness value.It is said differently, it is simply that the minimum luminance value of these pixels is non-normally low, it is possible to determine according to minimum value in a Color Channel RGB, centered by Ω (x), be positioned at a localized mass of x.
Finding 2 points that haze is the denseest and the lightest in the picture, wherein, described dark value is more little, represents concentration more big.Using this two-point drawing straight line as depth of field datum line, on this straight line, on the intersecting lens of predetermined angle, the dark value of pixel takes the dark average of all pixels on this intersecting lens, as environment channel value.It addition, take that the denseest pixel value of haze to be set to global context light AC, it refers to the environment light in air.Described predetermined angle, including 30 °, 60 °, 90 ° etc., does not limit.
Wherein in an embodiment, choose the average of the dark value of each pixel on the vertical line of described depth of field datum line, as the environment channel value of each pixel on described vertical line.Implement the present embodiment, it is possible to react the gradual change law of " haze " concentration better.Be conducive to obtaining apparent image.
So far, on present image, each pixel all has been assigned environment channel value, calculates according to formula (2) and obtains absorbance t (x)；
Wherein, 0 < ω≤1, default value is 0.95；The localized mass that Ω (x) is image block x.
According to described global context light value ACCalculate absorbance with the environment channel value of each pixel on present image, according to the described absorbance obtained, in conjunction with (3) formula, present image is carried out mist elimination process.
Wherein, I (x) is a physical model preset.
Wherein in the middle of an embodiment, utilize described absorbance, choose default atmospheric scattering physical model and present image is carried out mist elimination process.I.e. I (x)=J (x) t (x)+A (1-t (x)), the image that I (x) is an actually-received for greasy weather situation lower sensor, A (1-t (x)) is environment light, and A is global context light, t0=0.1.
In computer vision and computer graphics field, the atmospherical scattering model in greasy weather is generally as follows expression:
I (x)=J (x) t (x)+A (1-t (x)) (1-1)
Section 1 is direct attenuation components, and Section 2 A (1-t (x)) is environment light.The image that I (x) is an actually-received for greasy weather situation lower sensor, J (x) is scene radiance, A is global context light, and t (x) is absorbance, represents the radiometric ratio not eventually arrived at sensor in scene by KPT Scatter.
Under air is assumed uniformly, absorbance t (x) can be expressed as:
T (x)=e-βd(x)(1-2)
Wherein, β represents the scattering coefficient of air, and d represents that the depth of field, t (x) represent that scene radiance is along with the increase of the depth of field and exponential damping.
The mist elimination of Misty Image is exactly mainly from degraded image I (x) restoration scenario radiancy J (x).
To sum up, the present invention does not limit a certain physical model or a certain transmittance calculation mode.
Fig. 3 is the schematic diagram of image demister of the present invention, including:
Dark computing unit, for calculating the dark value of each pixel of present image；
Environment light acquiring unit, is used for the minima taking described dark value as global context light value；
Environment passage acquiring unit, for with the straight line at minimum and maximum two the pixel places of described dark value in present image for depth of field datum line, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
The mist elimination processing unit being respectively connected with described environment light acquiring unit, described environment passage acquiring unit, for calculating absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process.
Fig. 3 and Fig. 1 is corresponding, and in figure, the method for operation of unit is identical with method.
Fig. 4 is the embodiment schematic diagram of image defogging method of the present invention.
As described in Figure 4, the present invention, including:
The dark amending unit being connected with described dark computing unit, for traveling through each pixel according to preset order, according to comprise current detection pixel default localized mass in the described dark value of each pixel, choose the minimum dark value dark value as the pixel correction of described current detection.
Wherein in the middle of an embodiment, described passage amending unit, including:
Unit is chosen in region, for choosing centered by the pixel of described current detection, presets rectangular area long, default wide, as described localized mass.
Wherein in the middle of an embodiment, described environment passage acquiring unit, including:
Intersecting lens chooses unit, and the average of the dark value of each pixel on the vertical line choosing described depth of field datum line, as the environment channel value of each pixel on described vertical line.
Wherein in the middle of an embodiment, described mist elimination processing unit, including:
Unit chosen by model, is used for utilizing described absorbance, chooses default atmospheric scattering physical model and present image is carried out mist elimination process.
Fig. 4 and Fig. 2 is corresponding, and in figure, the method for operation of unit is identical with method.
Embodiment described above only have expressed the several embodiments of the present invention, and it describes comparatively concrete and detailed, but therefore can not be interpreted as the restriction to the scope of the claims of the present invention.It should be pointed out that, for the person of ordinary skill of the art, without departing from the inventive concept of the premise, it is also possible to making some deformation and improvement, these broadly fall into protection scope of the present invention.Therefore, the protection domain of patent of the present invention should be as the criterion with claims.
Claims (10)
1. an image defogging method, it is characterised in that including:
Calculate the dark value of each pixel of present image；Wherein, the minima of described dark value is taken as global context light value；
In present image, the straight line at two pixel places that described dark value is minimum and maximum is depth of field datum line, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
Calculate absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process；
According to expression formulaCalculate absorbance t (x), wherein 0 < ω≤1, c ∈ { localized mass that r, g, b}, Ω (x) are image block x, AcFor global context light value.
2. image defogging method according to claim 1, it is characterised in that after calculating the step of dark value of each pixel of present image, before taking the minima of the described dark value step as global context light value, including:
Travel through each pixel according to preset order, according to comprise current detection pixel default localized mass in the described dark value of each pixel, choose the minimum dark value dark value as the pixel correction of described current detection.
3. image defogging method according to claim 2, it is characterised in that:
Choose centered by the pixel of described current detection, preset rectangular area long, default wide, as described localized mass.
4. the image defogging method according to any one of claims 1 to 3, it is characterized in that, described basis and described depth of field datum line form the average of the dark value of each pixel on the intersecting lens of predetermined angle, as the step of the environment channel value of each pixel on described intersecting lens, including:
Choose the average of the dark value of each pixel on the vertical line of described depth of field datum line, as the environment channel value of each pixel on described vertical line.
5. image defogging method according to claim 1, it is characterised in that according to the described absorbance obtained, present image is carried out the step of mist elimination process, including:
Utilize described absorbance, choose default atmospheric scattering physical model and present image is carried out mist elimination process.
6. an image demister, it is characterised in that including:
Dark computing unit, for calculating the dark value of each pixel of present image；
Environment light acquiring unit, is used for the minima taking described dark value as global context light value；
Environment passage acquiring unit, for with the straight line at minimum and maximum two the pixel places of described dark value in present image for depth of field datum line, according to forming the average of the dark value of each pixel on the intersecting lens of predetermined angle with described depth of field datum line, as the environment channel value of each pixel on described intersecting lens；
The mist elimination processing unit being respectively connected with described environment light acquiring unit, described environment passage acquiring unit, for calculating absorbance according to the environment channel value of each pixel on described global context light value and present image, according to the described absorbance obtained, present image is carried out mist elimination process；
Described mist elimination processing unit is according to expression formulaCalculate absorbance t (x), wherein 0 < ω≤1, c ∈ { localized mass that r, g, b}, Ω (x) are image block x, AcFor global context light value.
7. image demister according to claim 6, it is characterised in that including:
The dark amending unit being connected with described dark computing unit, for after the dark value of each pixel of described dark computing unit calculating present image, each pixel is traveled through according to preset order, according to comprise current detection pixel default localized mass in the described dark value of each pixel, choose the minimum dark value dark value as the pixel correction of described current detection；
Described environment light acquiring unit, for described dark amending unit choose minimum dark value as the dark value of the pixel correction of described current detection after, take the minima of described dark value as global context light value.
8. image demister according to claim 7, it is characterised in that described dark amending unit, including:
Unit is chosen in region, for choosing centered by the pixel of described current detection, presets rectangular area long, default wide, as described localized mass.
9. the image demister according to any one of claim 6 to 8, it is characterised in that described environment passage acquiring unit, including:
Intersecting lens chooses unit, and the average of the dark value of each pixel on the vertical line choosing described depth of field datum line, as the environment channel value of each pixel on described vertical line.
10. image demister according to claim 6, it is characterised in that described mist elimination processing unit, including:
Unit chosen by model, is used for utilizing described absorbance, chooses default atmospheric scattering physical model and present image is carried out mist elimination process.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310349564.1A CN103426151B (en) | 2013-08-12 | 2013-08-12 | A kind of image defogging method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310349564.1A CN103426151B (en) | 2013-08-12 | 2013-08-12 | A kind of image defogging method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103426151A CN103426151A (en) | 2013-12-04 |
| CN103426151B true CN103426151B (en) | 2016-06-29 |
Family
ID=49650845
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310349564.1A Expired - Fee Related CN103426151B (en) | 2013-08-12 | 2013-08-12 | A kind of image defogging method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103426151B (en) |
Families Citing this family (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103824260A (en) * | 2013-12-20 | 2014-05-28 | 中国科学院遥感与数字地球研究所 | High-resolution remote sensing image mist rapid removal technology |
| CN103745438A (en) * | 2013-12-31 | 2014-04-23 | 中国科学院深圳先进技术研究院 | Haze removing method for large-area background light haze-containing image |
| CN104182943B (en) * | 2014-08-27 | 2015-12-02 | 湖南大学 | A kind of single image defogging method capable merging human-eye visual characteristic |
| CN104331866A (en) * | 2014-10-29 | 2015-02-04 | 中国科学院深圳先进技术研究院 | Image haze removal processing method and device |
| CN104766284B (en) * | 2015-04-17 | 2018-11-23 | 河海大学常州校区 | The restored method of underwater colored blur image |
| CN108681990B (en) * | 2018-04-04 | 2022-05-24 | 高明合 | Real-time haze early warning method and system |
| CN110493509B (en) * | 2019-06-26 | 2021-03-19 | 盐城华昱光电技术有限公司 | Image information processing system and method of laser module |
| CN110544213B (en) * | 2019-08-06 | 2023-06-13 | 天津大学 | Image defogging method based on global and local feature fusion |
| CN112864947B (en) * | 2021-03-02 | 2022-07-05 | 山东鲁软数字科技有限公司智慧能源分公司 | Visual monitoring system and method |
| CN113014773B (en) * | 2021-03-02 | 2023-08-01 | 山东鲁软数字科技有限公司智慧能源分公司 | Video visualization monitoring system and method for overhead line |
| CN113030638B (en) * | 2021-03-02 | 2022-06-14 | 山东鲁软数字科技有限公司智慧能源分公司 | Overhead line image visual monitoring system and method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063706A (en) * | 2010-12-23 | 2011-05-18 | 哈尔滨工业大学（威海） | Rapid defogging method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8340461B2 (en) * | 2010-02-01 | 2012-12-25 | Microsoft Corporation | Single image haze removal using dark channel priors |
-
2013
- 2013-08-12 CN CN201310349564.1A patent/CN103426151B/en not_active Expired - Fee Related
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102063706A (en) * | 2010-12-23 | 2011-05-18 | 哈尔滨工业大学（威海） | Rapid defogging method |
Non-Patent Citations (2)
| Title |
|---|
| Image Dehazing Algorithm Based on Atmosphere Scatters Approximation Model;Zhongyi Hu等;《Neural Information Processing》;20121231;第7667卷;159-168 * |
| 基于单幅图像景深和大气散射模型的去雾方法;郭珈等;《中国图象图形学报》;20120131;第17卷(第1期);27-32 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103426151A (en) | 2013-12-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103426151B (en) | A kind of image defogging method and device | |
| CN102831591B (en) | Gaussian filter-based real-time defogging method for single image | |
| CN103218778B (en) | The disposal route of a kind of image and video and device | |
| CN102411774B (en) | Processing method, device and system based on single-image defogging | |
| CN102063706B (en) | Rapid defogging method | |
| CN104036466B (en) | A kind of video defogging method and system | |
| CN103150708B (en) | Based on the image Quick demisting optimization method of black channel | |
| CN104794697B (en) | A kind of image defogging method based on dark primary priori | |
| CN102768760B (en) | Quick image dehazing method on basis of image textures | |
| CN103198459B (en) | Haze image rapid haze removal method | |
| CN104867121B (en) | Image Quick demisting method based on dark primary priori and Retinex theories | |
| CN105023256B (en) | A kind of image defogging method and system | |
| CN107301623A (en) | A kind of traffic image defogging method split based on dark and image and system | |
| CN103578083B (en) | Single image defogging method based on associating average drifting | |
| CN107578386A (en) | A kind of optimization defogging processing method of unmanned plane shooting image | |
| CN102750674A (en) | Video image defogging method based on self-adapting allowance | |
| CN104616258B (en) | A kind of rapid defogging method for road image | |
| CN103049888A (en) | Image/video demisting method based on combination of dark primary color of atmospheric scattered light | |
| CN103971337A (en) | Infrared image haze removal method based on atmospheric transmission characteristics | |
| CN102436643A (en) | Image defogging method facing to atmospheric scattering proximity effect | |
| CN105139347A (en) | Polarized image defogging method combined with dark channel prior principle | |
| CN105913390A (en) | Image defogging method and system | |
| CN103020914A (en) | Rapid image defogging method based on spatial continuity principle | |
| CN106780380A (en) | A kind of image defogging method and system | |
| CN104766286A (en) | Image defogging device and method based on pilotless automobile |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: 510670 Guangdong Province, Guangzhou high tech Industrial Development Zone Kezhu Road No. 233 Patentee after: VTRON GROUP Co.,Ltd. Address before: 510663 Guangzhou province high tech Industrial Development Zone, Guangdong, Cai road, No. 6, No. Patentee before: VTRON TECHNOLOGIES Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160629 Termination date: 20210812 |