CN111738938B - Nonuniform atomization video optimization method based on prior target identification - Google Patents
Nonuniform atomization video optimization method based on prior target identification Download PDFInfo
- Publication number
- CN111738938B CN111738938B CN202010487198.6A CN202010487198A CN111738938B CN 111738938 B CN111738938 B CN 111738938B CN 202010487198 A CN202010487198 A CN 202010487198A CN 111738938 B CN111738938 B CN 111738938B
- Authority
- CN
- China
- Prior art keywords
- video
- defogging
- image
- distribution function
- atmospheric
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 17
- 238000000889 atomisation Methods 0.000 title claims abstract description 12
- 238000005457 optimization Methods 0.000 title claims description 10
- 238000005286 illumination Methods 0.000 claims abstract description 24
- 238000005315 distribution function Methods 0.000 claims abstract description 21
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims abstract description 20
- 239000003595 mist Substances 0.000 claims abstract description 11
- 238000005507 spraying Methods 0.000 claims abstract description 8
- 239000011159 matrix material Substances 0.000 claims description 13
- 239000007921 spray Substances 0.000 claims description 7
- 238000002834 transmittance Methods 0.000 claims description 7
- 239000000779 smoke Substances 0.000 claims description 3
- 206010047571 Visual impairment Diseases 0.000 claims 1
- 238000007599 discharging Methods 0.000 claims 1
- 230000000694 effects Effects 0.000 abstract description 12
- 238000004140 cleaning Methods 0.000 abstract description 4
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000003467 diminishing effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a nonuniform atomization video enhancement method based on prior target identification. When water jet cleaning operation is carried out on an industrial site, water mist in the atmosphere can be unevenly diffused, video data obtained on the operation site have the characteristics of uneven illumination and atomization under the influence of atmospheric light reflection, and the existing defogging theory cannot achieve a good processing effect on videos. According to the invention, a priori target is formulated, template matching is carried out on a traditional defogging effect image, a fog spraying point is identified, a spatial distribution function and a defogging weight distribution function of an atmospheric illumination value are established according to the spraying point, the functions are dynamically updated according to the change of the water fog spraying point among different frames of the video, and a good video defogging effect is achieved.
Description
Technical Field
The invention belongs to the field of computer graphics and computer image processing, and particularly relates to a method for performing defogging optimization on a non-uniform atomization video affected by an artificial interference source.
Background
When the water jet cleaning operation is carried out in an industrial field, the water mist in the atmosphere can be unevenly diffused, and the video collected on the field is usually atomized under the influence of atmospheric light reflection. Besides, strong light source interference, smoke emission of equipment and the like can also cause similar effects, and defogging is one of important preprocessing flows in video analysis.
Dark channel prior defogging is a classical theory in the field of image defogging. In most non-sky images, at least one color channel value is low in most pixel points. The dark channel of the known image in the fog image forming model is calculated, the smaller item is discarded, meanwhile, the atmospheric illumination value in the image is counted, the atmospheric transmittance can be calculated, and the image defogging processing is completed.
The theory is suitable for defogging treatment of natural images. However, the defogging effect of the non-uniform fogging images and videos formed under the influence of artificial interference sources is not ideal. Therefore, a specific optimization method is needed to solve this problem.
Disclosure of Invention
In order to solve the technical problems in the prior art, the invention provides a non-uniform atomization video optimization method based on prior target identification. The invention is suitable for videos to be defogged, which are collected under the conditions of known artificial interference sources and non-uniform illumination, wherein the interference comprises but is not limited to: disturbance of intense light sources, disturbance of water mist spray, disturbance of smoke discharge, etc. are known.
The method mainly comprises the following steps:
(1) obtaining a preliminary defogging image on the basis of a dark channel prior defogging theory;
(2) carrying out template matching on the traditional defogging effect image by formulating a prior target, and identifying a fog spraying point;
(3) establishing a spatial distribution function and a defogging weight distribution function of the atmospheric illumination value according to the ejection point;
(4) and dynamically updating the function according to the change of the water mist spraying points among different frames of the video, and outputting an optimized video.
In step (1), it is necessary to use a recognized fog pattern forming model:
I(x)=J(x)t(x)+A(1-t(x)) (1)
wherein A is an atmospheric illumination value, t (x) is a transmittance, J (x) is a real image, and I (x) is a fogging image. To calculate the real image j (x), the atmospheric illumination value a and the transmittance t (x) need to be obtained. The method for obtaining the A value is to calculate the dark channel of the image, take the position of the first 1 per mill bright spot, and calculate the average value of the gray scale of the corresponding position in the atomization image. The formula for obtaining the transmittance t (x) is:
wherein Ω (x) is a sliding window on the image, the minimum value of the dark channel in the window, the atmospheric illumination value a, and the defogging weight w are obtained, the transmittance estimated value t (x) in the window can be calculated, and finally the real image j (x) is calculated according to the formula (1).
In the step (2), firstly, a priori target, namely an image of an artificial interference source, needs to be formulated, and the size of the image is smaller than that of the image of the atomization video. And (3) carrying out template matching on the atomized image by calculating normalized square difference by using a sliding window strategy. And obtaining the coordinate of the point with the minimum square difference in the return matrix, namely the coordinate position of the interference source in the image.
In the step (3), the established atmosphere illumination distribution function and defogging weight distribution function are different according to different interference sources. For example, the light divergence of a strong light source can affect the overall picture quality of a video, and the light intensity of the strong light source has a determined model; the spray mist affects only the local area of the atmospheric illumination (reflected light), with the white field at the center of the spray being the most intense and then rapidly diminishing. The distribution function constructed in this example for the case of spraying water mist is as follows:
wherein A (i, j) is any element in the atmosphere illumination value distribution function matrix, and the A value in the step (1) is expanded into a matrix with the same size as the image; α is the atmospheric illuminance ratio (ratio of atmospheric illumination at the edge of the image to atmospheric illumination at the prior object); a. the 0 Is the image average atmospheric illumination value; (x) 0 ,y 0 ) Obtaining prior target coordinates for template matching; (h) a 0 ,w 0 ) Is the image size; w (i, j) is an arbitrary element in the defogging weight distribution function matrix, and the W value in step (1) is also generalized to a weight matrix of the same size as the image. According to A, W matrix and formula (2), the optimized perspective ratio t (x) can be calculated, and then according to formula (1), the real image J (x) can be calculated.
In the step (4), firstly, the J (x) image is enhanced, and the contrast and the brightness of the image are improved; after each frame of image is processed, the preliminary defogging of the next frame of image can be carried out. Before that, the A, W matrix is updated, so that better preliminary defogging effect can be achieved. If the position of the artificial interference source in the video does not move violently, namely the element value of A, W is relatively stable, the preliminary defogging (step 1) can be omitted, and the template matching (step 2) is directly carried out.
The invention has the beneficial effects that:
1. the method is suitable for processing the non-uniform atomization video in the industrial field, has a good inhibition effect on the atomization influence of the artificial interference source, and is very important for subsequent video analysis.
2. According to different scenes, the method can change the atmospheric illumination distribution function and the defogging weight distribution function, can be popularized to other application environments, and has good subsequent expansibility.
Drawings
FIG. 1 is a flow chart of the main algorithm of the present invention.
Fig. 2 is an original video screenshot used by the present invention.
Fig. 3 is a video screenshot of the preliminary defogging process performed in step (1) according to the present invention.
Fig. 4 is a final video screenshot after processing by the present invention.
Detailed Description
The invention is further described below with reference to the accompanying drawings.
As shown in fig. 1, the main flow of the present invention is shown, and the specific implementation manner is described above.
The original video adopted by the embodiment is from an industrial site where a high-pressure water derusting robot cleans a wall surface, as shown in fig. 2. When a high-pressure water spray rod carried by the robot carries out cleaning operation, a large amount of water mist is formed around the robot, so that the judgment on the cleaning effect of the robot is influenced, and therefore, the defogging treatment is required to be carried out on a video.
In the step (1), firstly, obtaining information such as video size, frame rate and the like, taking a first frame of a video as a source image, traversing the image to obtain a dark channel image of the image, and filtering the dark channel image with a minimum value in a small range; then, counting the position of the front 1 per thousand brightness points in the dark channel image, and calculating the average gray level of the pixel points at the position in the source image to be used as an atmospheric illumination value A; setting the initial defogging weight value w to be 0.92, and calculating the transmissivity t (x) of each pixel point of the image according to a formula (2); finally, calculating a preliminary defogging image J (x) according to a formula (1); in this embodiment, the preliminary defogging video screenshot is as shown in fig. 3, and the effect is not ideal.
In the step (2), the upper half part of the high-pressure water spray boom is taken as a prior target, template matching is carried out on the preliminary defogged image, and the coordinate of the high-pressure water spray boom in the image is obtained; the center of the water jet is positioned at the upper right part of the prior target image and can be obtained by coordinate conversion; the result of the template matching is shown as a rectangular box in fig. 4.
In the step (3), according to the characteristics of the embodiment, a large number of 'white field' bright spots are gathered at the center of the water jet, the density of the water mist is rapidly reduced along with the increase of the spatial distance, and the edge of the image is basically not affected by the water mist. According to statistics, the gray values of 'white field' bright spots in the image are all 255, the atmospheric illumination ratio is about 0.8, and an atmospheric illumination distribution function is constructed as shown in a formula (3); because the defogging effect needs to be enhanced at the center of the jet water flow, as the space distance is increased, the defogging demand is gradually weakened, and a defogging weight distribution function is constructed as shown in a formula (4); and (3) substituting the A, W matrix into a formula (2), calculating the optimized transmittance t (x), and substituting the parameters into a formula (1) to obtain a real image J (x).
In the step (4), because the image after defogging is dark, for the embodiment, the contrast is improved by 1.4 times, the brightness is increased by 10 gray levels, the final image is written into the video, and then the next frame is processed; the A, W matrix is updated to the value of the previous frame before the preliminary defogging of the next frame, so that a better preliminary defogging effect can be obtained.
In this embodiment, the final processed video screenshot is shown in fig. 4.
Claims (5)
1. A nonuniform atomization video optimization method based on prior target identification is characterized by comprising the following steps:
the method comprises the following steps of (1) obtaining a preliminary defogging image on the basis of a dark channel prior defogging theory;
step (2) template matching is carried out on the preliminary defogging image by formulating a prior target, and a fog spraying point is identified;
step (3) establishing a spatial distribution function and a defogging weight distribution function of the atmospheric illumination value according to the ejection point;
dynamically updating the function according to the change of water mist spraying points among different frames of the video, and outputting an optimized video;
the spatial distribution function and the defogging weight distribution function of the atmospheric illumination value in the step (3) are functions which are formed by judging the non-uniform atomization degree according to the requirements of a video scene or by using a distribution model of an interference source for reference;
the spatial distribution function of the atmospheric illumination values is:
the defogging weight distribution function is as follows:
wherein A (i, j) is any element in the atmosphere illumination value distribution function matrix; alpha is the atmospheric illuminance ratio; a. the 0 Is the image average atmospheric illumination value; (x) 0 ,y 0 ) Obtaining prior target coordinates for template matching; (h) 0 ,w 0 ) Is the image size; w (i, j) is an arbitrary element in the defogging weight distribution function matrix.
2. The method of a priori object recognition based non-uniform fogging video optimization of claim 1, wherein: the dark channel prior defogging theory in the step (1) generates a defogged image theory by calculating an image dark channel, counting an atmospheric illumination value and calculating atmospheric transmittance.
3. The method of a priori object recognition based non-uniform fogging video optimization of claim 1, wherein: and (3) the prior target is established by manually marking a known video interference source, and the image size of the prior target is smaller than that of the video image.
4. The method of a priori object recognition based non-uniform fogging video optimization of claim 3, wherein: such known sources of visual disturbance include light sources emitting intense light, spray heads emitting water mist and/or device outlets discharging smoke, etc.
5. The method for non-uniform fogging video optimization based on a priori object identification as claimed in claim 1, wherein: the dynamic updating of the function refers to updating (x) after identifying the prior target in each frame of image 0 ,y 0 ) Sitting chairAnd marking values, and updating the matrix A and the matrix W according to the functions.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010487198.6A CN111738938B (en) | 2020-06-01 | 2020-06-01 | Nonuniform atomization video optimization method based on prior target identification |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010487198.6A CN111738938B (en) | 2020-06-01 | 2020-06-01 | Nonuniform atomization video optimization method based on prior target identification |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111738938A CN111738938A (en) | 2020-10-02 |
CN111738938B true CN111738938B (en) | 2022-09-09 |
Family
ID=72646632
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010487198.6A Active CN111738938B (en) | 2020-06-01 | 2020-06-01 | Nonuniform atomization video optimization method based on prior target identification |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111738938B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104299192A (en) * | 2014-09-28 | 2015-01-21 | 北京联合大学 | Single image defogging method based on atmosphere light scattering physical model |
CN106157267A (en) * | 2016-07-12 | 2016-11-23 | 中国科学技术大学 | A kind of image mist elimination absorbance optimization method based on dark channel prior |
CN106251296A (en) * | 2016-07-06 | 2016-12-21 | 北京元心科技有限公司 | A kind of image defogging method and system |
CN108717686A (en) * | 2018-04-04 | 2018-10-30 | 华南理工大学 | A kind of real-time video defogging method based on dark channel prior |
CN110288539A (en) * | 2019-06-03 | 2019-09-27 | 哈尔滨工程大学 | A kind of mobile clear method of underwater picture with dark channel prior in color combining space |
CN110544220A (en) * | 2019-09-05 | 2019-12-06 | 北京天地玛珂电液控制***有限公司 | intelligent defogging method, storage medium and system for underground video image of coal mine |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2013018101A1 (en) * | 2011-08-03 | 2013-02-07 | Indian Institute Of Technology, Kharagpur | Method and system for removal of fog, mist or haze from images and videos |
-
2020
- 2020-06-01 CN CN202010487198.6A patent/CN111738938B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104299192A (en) * | 2014-09-28 | 2015-01-21 | 北京联合大学 | Single image defogging method based on atmosphere light scattering physical model |
CN106251296A (en) * | 2016-07-06 | 2016-12-21 | 北京元心科技有限公司 | A kind of image defogging method and system |
CN106157267A (en) * | 2016-07-12 | 2016-11-23 | 中国科学技术大学 | A kind of image mist elimination absorbance optimization method based on dark channel prior |
CN108717686A (en) * | 2018-04-04 | 2018-10-30 | 华南理工大学 | A kind of real-time video defogging method based on dark channel prior |
CN110288539A (en) * | 2019-06-03 | 2019-09-27 | 哈尔滨工程大学 | A kind of mobile clear method of underwater picture with dark channel prior in color combining space |
CN110544220A (en) * | 2019-09-05 | 2019-12-06 | 北京天地玛珂电液控制***有限公司 | intelligent defogging method, storage medium and system for underground video image of coal mine |
Non-Patent Citations (8)
Title |
---|
A fast video image defogging algorithm based on dark channel prior;Erhu Zhang, et.al;《2013 6th International Congress on Image and Signal Processing (CISP)》;IEEE;20140220;第1卷;全文 * |
Research on Image Enhancement Algorithm of Coal Mine Dust;Dongmei Wu, et.al;《2018 International Conference on Sensor Networks and Signal Processing (SNSP)》;IEEE;20190117;全文 * |
UNDERWATER IMAGE ENHANCEMENT BY WAVELENGTH COMPENSATION AND DEHAZING;Hohn Y. Chiang, et. al;《IEEE TRANSACTIONS ON IMAGE PROCESSING》;IEEE;20120430;第21卷(第4期);全文 * |
一种暗原色先验的图像去雾改进算法;宋东辉等;《应用科技》;20161231;第43卷(第6期);全文 * |
人工光源条件下夜间雾天图像建模及去雾;杨爱萍等;《电子与信息学报》;20180630;第40卷(第6期);第4页 * |
光照模型的分布函数及其评估;邹荣金;《计算机工程》;19950531;第21卷;全文 * |
基于暗通道先验的视频去雾算法;贾银亮等;《电子测量技术》;20181031;第41卷(第20期);第2页 * |
雾状水幕2种常用分布函数的近似转换;陈中伟等;《清华大学学报(自然科学版)》;20110430;第51卷(第4期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN111738938A (en) | 2020-10-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Zhang et al. | Fast haze removal for nighttime image using maximum reflectance prior | |
JP4629364B2 (en) | A method for adaptively updating a background image representing the background of a scene | |
CN107103591B (en) | Single image defogging method based on image haze concentration estimation | |
CN108765342A (en) | A kind of underwater image restoration method based on improvement dark | |
CN106952280B (en) | A kind of spray gun paint amount uniformity detection method based on computer vision | |
CN111145105B (en) | Image rapid defogging method and device, terminal and storage medium | |
CN109191429B (en) | 3D printing nozzle detection method based on machine vision | |
CN103020921A (en) | Single image defogging method based on local statistical information | |
CN106296618A (en) | A kind of color image defogging method based on Gaussian function weighted histogram regulation | |
CN111598814B (en) | Single image defogging method based on extreme scattering channel | |
CN105488769B (en) | A kind of real-time video defogging method | |
Mei et al. | Single image dehazing using dark channel fusion and haze density weight | |
CN109003238A (en) | A kind of image haze minimizing technology based on model and histogram and grey level enhancement | |
CN110349113B (en) | Adaptive image defogging method based on dark primary color priori improvement | |
CN114862897A (en) | Image background processing method and device and electronic equipment | |
CN111738938B (en) | Nonuniform atomization video optimization method based on prior target identification | |
CN105913391B (en) | A kind of defogging method can be changed Morphological Reconstruction based on shape | |
CN105303529A (en) | Digital image haze removing method based on histogram equalization | |
CN111738934B (en) | Automatic red eye repairing method based on MTCNN | |
CN113379631A (en) | Image defogging method and device | |
Sun et al. | Fusion algorithm for foggy image enhancement based on transmittance weight factor | |
CN111339968A (en) | Method for acquiring two-dimensional coordinates of indoor high-density aquaculture fish based on computer vision technology | |
Ganguly et al. | Fusion of Mathematical Morphology with Adaptive Gamma Correction for Dehazing and Visibility Enhancement of Images | |
CN110232709A (en) | A kind of line-structured light Light stripes center extraction method becoming Threshold segmentation | |
CN108133461A (en) | A kind of real-time video low-light (level) Enhancement Method |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |