CN113920136A - Improved dark channel prior defogging algorithm - Google Patents
Improved dark channel prior defogging algorithm Download PDFInfo
- Publication number
- CN113920136A CN113920136A CN202111191635.0A CN202111191635A CN113920136A CN 113920136 A CN113920136 A CN 113920136A CN 202111191635 A CN202111191635 A CN 202111191635A CN 113920136 A CN113920136 A CN 113920136A
- Authority
- CN
- China
- Prior art keywords
- dark channel
- channel
- image
- algorithm
- improved
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 claims abstract description 37
- 238000001914 filtration Methods 0.000 claims abstract description 20
- 238000002834 transmittance Methods 0.000 claims abstract description 18
- 238000012545 processing Methods 0.000 claims abstract description 16
- 238000004364 calculation method Methods 0.000 claims abstract description 13
- 230000011218 segmentation Effects 0.000 claims abstract description 4
- 230000000694 effects Effects 0.000 claims description 12
- 125000001475 halogen functional group Chemical group 0.000 claims description 10
- 238000005070 sampling Methods 0.000 claims description 4
- 238000004971 IR microspectroscopy Methods 0.000 claims description 3
- 238000012060 immune response imaging Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 5
- 238000013135 deep learning Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 241000270295 Serpentes Species 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000004927 fusion Effects 0.000 description 2
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000000903 blocking effect Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The invention discloses an improved dark channel prior defogging algorithm, which comprises the following steps: using Rayleigh scattering theory to improve the estimated value of atmospheric light; a segmentation method based on an active contour model is adopted to segment a sky region to obtain a sky region and a non-sky region; estimating dark channel values using different sizes of filtering; the calculation of the transmittance is simplified, and the processing speed of the algorithm is improved. According to the method, firstly, on the basis of an original dark channel prior defogging algorithm, a Rayleigh scattering theory is added to improve an atmospheric light estimation method, then a fog image is divided into a sky region and a non-sky region, dark channel images with different filter sizes are fused to obtain a fused dark channel, and finally, the solving process of the transmissivity is simplified to improve the operation speed of the algorithm and enhance the real-time property.
Description
Technical Field
The invention belongs to the technical field of image defogging, and particularly relates to an improved dark channel prior defogging algorithm.
Background
Outdoor images photographed in environments such as fog and haze have a large number of suspended particles in the atmosphere, and the image quality is affected to various degrees in terms of contrast, color, and the like. However, in many image processing systems such as target detection and satellite remote sensing, people have strict requirements on the quality of input images. Therefore, the method for carrying out the sharpening processing on the foggy day image with the degraded quality has very practical significance in the field of computer vision.
At the initial stage of image defogging work, methods such as improving image contrast and saturation are often adopted to represent main characteristics of an image, and the method can achieve the purpose of image clearness, but is low in reliability and prone to large-scale information loss. With the continuous development of modern science and technology and the continuous improvement of image quality requirements, the atmospheric scattering model is applied to the field of image defogging, and a large number of defogging algorithms depending on the atmospheric scattering model appear, and the algorithms fully consider the image degradation process and physically remove the fog existing in the image. At present, along with the rise of deep learning, processing a foggy image by using a deep learning method gradually becomes a mainstream, but the method is limited by many factors, and the problem of incomplete defogging still exists when the image in a real scene is subjected to sharpening processing.
In summary, the method for performing the sharpening process on the foggy day image can be divided into three categories, namely an algorithm based on image enhancement, an algorithm based on deep learning and a defogging algorithm based on a physical model. The defogging algorithm based on image enhancement can effectively improve the contrast and color information of an image, but can sacrifice part of detail information, such as the defogging algorithm based on histogram equalization, curvelet transformation and Retinex. Image defogging algorithms based on deep learning have emerged in the recent years. Ren et al propose a convolutional neural network for multi-scale feature extraction and defogging by estimating fog map transmittance. But often perform less than fully when dealing with fog patterns in real weather. The defogging algorithm based on the physical model analyzes the process of image quality degradation, and the pertinence is stronger. Jackson et al propose a defogging method for estimating an initial transmittance graph by using Rayleigh scattering and establishing a scattering coefficient model and optimizing the transmittance graph by using guide filtering, thereby improving the operation speed but not completely defogging the image. He and the like propose a dark channel defogging algorithm, most of foggy images can be effectively recovered by using the algorithm, but the processed sky area can generate a serious halo effect because the bright sky area does not meet dark channel prior. Then He and the like propose to use guiding filtering to replace a soft matting method to optimize the transmissivity on the basis of a Dark Channel Prior (DCP) theory, so that the defogging efficiency is improved, but the problems of incomplete defogging, low processing speed and the like still exist after the processing. Aiming at the defects of methods such as He and the like, Liu and the like propose that a self-adaptive threshold value is introduced in the process of acquiring the dark primary color, so that the accuracy of the transmissivity is improved, but the operation efficiency is still lower.
Disclosure of Invention
Based on the defects of the prior art, the invention aims to provide an improved dark channel prior defogging algorithm to solve the problems of low defogging speed, color distortion and incomplete defogging of the traditional defogging method.
In order to solve the technical problems, the invention is realized by the following technical scheme:
the invention provides an improved dark channel prior defogging algorithm, which comprises the following steps:
s1: using Rayleigh scattering theory to improve the estimated value of atmospheric light;
s2: a segmentation method based on an active contour model is adopted to segment a sky region to obtain a sky region and a non-sky region;
s3: estimating dark channel values using different sizes of filtering;
s4: the calculation of the transmittance is simplified, and the processing speed of the algorithm is improved.
Optionally, the step S1 includes the following steps:
s11, according to Rayleigh scattering, selecting the pixel sequence with the highest intensity from the blue channel of the input image according to the column, and averaging the sequence data to obtain A1For eliminating the halo effect, A1The expression of (a) is as follows:
A1＝avg(max(Iin the B column direction))
In the formula: a. the1Average maximum of blue channel of fog pattern, IBIs the blue channel value of the fog map;
s12, calculating the average value A of the lowest pixels in the red and green channels2Further eliminate A1Later unnecessary texture information:
A2＝avg(min(IG)+min(IR))
in the formula: i isGIs the green channel value of the fog map, IRIs the red channel value of the fog map, A2Is the average of the darkest green channel and the darkest red channel;
s13, calculation A1And A2To estimate the atmospheric light a:
A＝A1+A2。
optionally, in step S2, the dark channel maps with different filter sizes are fused, and the edge area emphasizes using the dark channel map with a smaller filter size, so as to better extract the scene structure information of the depth discontinuity area and avoid halo artifacts from being generated in the edge area; the smooth area uses a larger filtering size dark channel image to ensure that the restored image does not generate color cast due to over-enhancement.
Further, the step S3 includes the following steps:
s31, converting RGB into a gray histogram;
s32, down-sampling the original image and reducing the original image into 1/4 of the original image;
and S33, acquiring the transmittance of the original image by adopting an interpolation method.
Therefore, the method is characterized in that firstly, on the basis of the original dark channel prior defogging algorithm, the Rayleigh scattering theory is added to improve the atmospheric light estimation method, then the fog image is divided into a sky region and a non-sky region, the dark channel images with different filter sizes are fused to obtain a fused dark channel, and finally, the solving process of the transmissivity is simplified to improve the operation speed of the algorithm and enhance the real-time property.
The foregoing description is only an overview of the technical solutions of the present invention, and in order to make the technical means of the present invention more clearly understood, the present invention may be implemented in accordance with the content of the description, and in order to make the above and other objects, features, and advantages of the present invention more clearly understood, the following detailed description is given in conjunction with the preferred embodiments, together with the accompanying drawings.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments will be briefly described below.
FIG. 1 is a flow chart of the improved dark channel a priori defogging algorithm of the present invention.
Detailed Description
Other aspects, features and advantages of the present invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawings, which form a part of this specification, and which illustrate, by way of example, the principles of the invention. In the referenced drawings, the same or similar components in different drawings are denoted by the same reference numerals.
In order to solve the problems of low defogging speed, color distortion, incomplete defogging and the like of the traditional defogging method, the invention provides an improved dark channel prior defogging algorithm, and the algorithm flow is shown in fig. 1. The improved dark channel prior defogging algorithm of the invention comprises the following steps:
s1: using Rayleigh scattering theory to improve the estimated value of atmospheric light;
s2: a segmentation method based on an active contour model is adopted to segment a sky region to obtain a sky region and a non-sky region;
s3: estimating dark channel values using different sizes of filtering;
s4: the calculation of the transmittance is simplified, and the processing speed of the algorithm is improved.
The atmospheric light value is an important parameter for image defogging, and the defogging performance of the whole image is reduced due to large calculation amount or large error. Conventional defogging algorithms estimate a by selecting the largest pixel. In order to accurately estimate the brightness of the pixel to express the haze density, the characteristic that blue light is easier to scatter than light of other colors is utilized by Rayleigh scattering theory, the B (blue) channel of the RGB three-color channel of the fog image is taken as the maximum value, and the R (red) and G (green) channels are respectively taken as the minimum value. An improved atmospheric light calculation method is thus proposed, incorporating both the brightest and darkest pixels into the calculation. The method mainly comprises the following steps:
(1) blue light is scattered most in RGB images according to rayleigh scattering. In order to comprehensively consider image parts with different depths, a pixel sequence with the highest intensity is selected from a blue channel of an input image in a row and an average value of sequence data is obtained to obtain A1For eliminating the halo effect, A1The expression of (a) is as follows:
A1＝avg(max(Iin the B column direction))
In the formula: a. the1Average maximum of blue channel of fog pattern, IBIs the blue channel value of the fog map.
(2) Calculate the average value A of the lowest pixels in the red and green channels2Further eliminate A1And then unnecessary texture information.
A2＝avg(min(IG)+min(IR))
In the formula: i isGIs the green channel value of the fog map, IRIs the red channel value of the fog map, A2The average of the darkest green channel and the darkest red channel.
(3) Calculation of A1And A2To estimate the atmospheric light A
A＝A1+A2。
An Active Con-road Model (ACM) method is a very representative method. The ACM combines the constrained information on the image with a priori knowledge to define a closed and continuous energy curve. The energy curve is continuously close to the target contour under the combined action of the internal force and the external force, and when the energy of the curve is minimum, the energy curve is considered to reach the target contour, namely the target edge is obtained. First, a curve is defined, x(s) ═ x(s), y (s)), s ∈ [0,1 ]. The curve energy function is:
in the formula, EintIs the energy term inside the curve, EextIs a curve external energy item, which is specifically expressed as:
alpha(s) and beta(s) are weight coefficients; alpha(s) | X'(s) | non-woven2The tension of the curve is controlled and the continuity of the curve is ensured for the elastic energy of the curve; beta(s) | X '(s)' non-woven phosphor2The curve rigidity energy controls the curve bending degree and ensures the curve smoothness. ^ represents gradient operator, I (x, y) represents image, GσGaussian filter operator with standard deviation sigma, convolution, EextIs given an energy constraint according to the image.
The GVF-Snake model introduces a GVF field V (x, y) ([ mu (x, y), ν (x, y)) on the basis of the Snake model, and the GVF field is used as a new external force field to replace a Gaussian external force field of the Snake model, so that the GVF-Snake model energy functional is obtained as follows,
μ is a parameter for controlling the integrand, and is set according to the amount of noise present in the image, and generally, the larger the noise, the larger the value. Variable ux、uy、vx、vyIs the first derivative of u, v in x, y direction. Term 1 is a smoothing energy term that can produce a slowly diffusing vector field; item 2 is the edge energy item, V ═ f at the edge of the image, such that the gradient forces are directed towards the edge, pushing the curve towards the edge, whereas | f | tends to 0 away from the edge, item 2 is | +, f |, thThe 1 term plays a major role in diffusing edge information toward a smooth region, so that the capture range of the contour curve increases.
In using the minimum filtering operation to obtain the dark channel, the filter size is a key parameter that affects the final defogging effect. The larger filter size is more compatible with the dark channel theory, because more dark pixels are included, but at the same time, the halo effect is caused by the large size filtering; conversely, a smaller filter size can effectively suppress halo effects in the edge region, but can cause oversaturation of the recovered image.
In order to solve the above defects, the invention provides an improved dark channel estimation method, which fuses dark channel images with different filtering sizes, and the edge area emphasizes the use of the dark channel image with a smaller filtering size, so as to better extract the scene structure information of the depth discontinuous area and avoid halo artifacts generated in the edge area; the smooth area uses a larger filtering size dark channel image to ensure that the restored image does not generate color cast due to over-enhancement. Therefore, the key of the fusion is to construct a weight map showing the depth discontinuity areas, and the median filtering technique, as a nonlinear filtering operation, can effectively suppress the impulse noise component while preserving edge information, which can be expressed as:
in the formula, omega1(x) The filter area is centered at x (taking a 15 × 15 square matrix), mean is the median filter operation, and W is the minimum channel of the input image.
Taking the image edge information image as a weight image, fusing the dark channel images with different sizes by adopting a pixel level fusion mode to obtain a fused dark channel Ifdark(x) The expression is:
in the formula (I), the compound is shown in the specification,
for filtering dark channel images of size 3 x 3,
is a dark channel image with a filter size of 15 x 15.
The algorithm of the invention utilizes the dark channel with smaller filtering size to effectively compensate the edge area of the dark channel, and utilizes the fog-free image recovered by the fused dark channel to avoid the supersaturation problem caused by the dark channel with small size and inhibit the halo effect caused by the dark channel with larger size to a certain extent.
Because the graph of the transmittance of the dark channel defogging is finer than other algorithms, and a large amount of processing time is occupied in the whole algorithm, and the processing efficiency of the program is reduced, if the precision is reduced a little properly, the defogging effect of the dark channel defogging should not be greatly different theoretically, the method adopted by the invention does not calculate the transmittance of the original image, but simplifies the calculation of the transmittance by the following method:
1. converting RGB into a gray level histogram;
2. down-sampling the original image, reducing the original image to 1/4 of the original image, storing in V (x), and solving for the transmittance of V (x);
3. and obtaining the approximate transmittance of the original image by adopting an interpolation method.
Experiments prove that the method greatly improves the execution speed, and the defogging effect is basically consistent with the original scheme. However, if the reduction factor is not particularly large, for example, 0.5 times the original size, the time consumption of the two scaling operations may also offset the profit gained by computing the transmittance map of the small map, and therefore this down-sampling rate must be chosen appropriately.
In the research, the traditional dark channel prior defogging has poor treatment effect on the sky part, and the sky often has large-area textures and blocking phenomena. The fundamental reason is that the sky part basically does not accord with the defogging prior condition of the dark channel, and the invention adopts a method for avoiding the excessive enhancement of the sky part in the image defogging algorithm, wherein (1) the image is converted into the gray scale, and the decolorizing algorithm with the contrast retaining function is used for retaining more edge information; (2) solving gradient information of the gray level image, and realizing by adopting an edge detection operator; (3) denoising and filtering the gradient information; (4) distinguishing gradient information according to a set gradient threshold and a brightness threshold; (5) and performing Gaussian feathering processing on the distinguished images.
After the sky area is obtained, the transmittance map of the sky area is uniformly set to a fixed value, but this is not good, and appropriate correction should be made according to a specific value. In the above operation with He, the resulting sky region does not really belong to the sky, and a certain point does not necessarily belong to the sky completely or not at all. Therefore, the present invention proposes an optimization method, as follows:
Jx＝(tp×Ix+Jx×(255-Ix))/255
wherein JxPoints referring to dark channels, tpThe fixed transmittance value similar to He is specified by adjusting the image processing effect, and in the above formula, if i (x) is 255, that is, completely belongs to the sky, the transmittance at that point is a fixed value, and if i (x) is 0, that is, completely does not belong to the sky, the value of the calculation formula is not changed, and normal defogging is not affected.
In addition, regarding the calculation of the atmospheric light value a, He proposes that the average value of the obtained pixels of the sky part is taken as a, which is also very reasonable, but in the actual processing, for some images without the sky part at all, the detected sky area is very small (obviously belonging to false detection, but the program does not know), and this is also unreasonable as the atmospheric light value. Therefore, the processing mode of the invention is to calculate the A value of the sky part and then detect the proportion of the sky pixel in the whole image.
While the foregoing is directed to the preferred embodiment of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
Claims (4)
1. An improved dark channel a priori defogging algorithm, comprising the steps of:
s1: using Rayleigh scattering theory to improve the estimated value of atmospheric light;
s2: a segmentation method based on an active contour model is adopted to segment a sky region to obtain a sky region and a non-sky region;
s3: estimating dark channel values using different sizes of filtering;
s4: the calculation of the transmittance is simplified, and the processing speed of the algorithm is improved.
2. The improved dark channel a priori defogging algorithm according to claim 1, wherein said step S1 comprises the steps of:
s11, according to Rayleigh scattering, selecting the pixel sequence with the highest intensity from the blue channel of the input image according to the column, and averaging the sequence data to obtain A1For eliminating the halo effect, A1The expression of (a) is as follows:
A1＝avg(max(Iin the B column direction))
In the formula: a. the1Average maximum of blue channel of fog pattern, IBIs the blue channel value of the fog map;
s12, calculating the average value A of the lowest pixels in the red and green channels2Further eliminate A1Later unnecessary texture information:
A2＝avg(min(IG)+min(IR))
in the formula: i isGIs the green channel value of the fog map, IRIs the red channel value of the fog map, A2Is the average of the darkest green channel and the darkest red channel;
s13, calculation A1And A2To estimate the atmospheric light a:
A＝A1+A2。
3. the improved dark channel a priori defogging algorithm according to claim 1, wherein in said step S2, the dark channel maps with different filtering sizes are fused, and the edge region emphasizes the use of the dark channel map with smaller filtering size so as to better extract the scene structure information of the depth discontinuity region and avoid the halo artifact generated in the edge region; the smooth area uses a larger filtering size dark channel image to ensure that the restored image does not generate color cast due to over-enhancement.
4. The improved dark channel a priori defogging algorithm according to claim 1, wherein said step S3 comprises the steps of:
s31, converting RGB into a gray histogram;
s32, down-sampling the original image and reducing the original image into 1/4 of the original image;
and S33, acquiring the transmittance of the original image by adopting an interpolation method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111191635.0A CN113920136A (en) | 2021-10-13 | 2021-10-13 | Improved dark channel prior defogging algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111191635.0A CN113920136A (en) | 2021-10-13 | 2021-10-13 | Improved dark channel prior defogging algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113920136A true CN113920136A (en) | 2022-01-11 |
Family
ID=79240100
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111191635.0A Pending CN113920136A (en) | 2021-10-13 | 2021-10-13 | Improved dark channel prior defogging algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113920136A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114155173A (en) * | 2022-02-10 | 2022-03-08 | 山东信通电子股份有限公司 | Image defogging method and device and nonvolatile storage medium |
-
2021
- 2021-10-13 CN CN202111191635.0A patent/CN113920136A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114155173A (en) * | 2022-02-10 | 2022-03-08 | 山东信通电子股份有限公司 | Image defogging method and device and nonvolatile storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106157267B (en) | Image defogging transmissivity optimization method based on dark channel prior | |
| CN107301623B (en) | Traffic image defogging method and system based on dark channel and image segmentation | |
| CN107527332A (en) | Enhancement Method is kept based on the low-light (level) image color for improving Retinex | |
| CN108389175B (en) | Image defogging method integrating variation function and color attenuation prior | |
| CN109255759B (en) | Image defogging method based on sky segmentation and transmissivity self-adaptive correction | |
| Gao et al. | Sand-dust image restoration based on reversing the blue channel prior | |
| CN112419162A (en) | Image defogging method and device, electronic equipment and readable storage medium | |
| CN114331873B (en) | Non-uniform illumination color image correction method based on region division | |
| CN107705258B (en) | Underwater image enhancement method based on three-primary-color combined pre-equalization and deblurring | |
| CN108133462B (en) | Single image restoration method based on gradient field region segmentation | |
| CN110827221A (en) | Single image defogging method based on double-channel prior and side window guide filtering | |
| CN111861896A (en) | UUV-oriented underwater image color compensation and recovery method | |
| CN109118440B (en) | Single image defogging method based on transmissivity fusion and adaptive atmospheric light estimation | |
| CN114693548B (en) | Dark channel defogging method based on bright area detection | |
| CN109345479B (en) | Real-time preprocessing method and storage medium for video monitoring data | |
| CN111325688A (en) | Unmanned aerial vehicle image defogging method fusing morphological clustering and optimizing atmospheric light | |
| CN113920136A (en) | Improved dark channel prior defogging algorithm | |
| CN113034379A (en) | Weather-time self-adaptive rapid image sharpening processing method | |
| CN112825189B (en) | Image defogging method and related equipment | |
| CN109636735B (en) | Rapid video defogging method based on space-time consistency constraint | |
| CN116862809A (en) | Image enhancement method under low exposure condition | |
| CN108765310B (en) | Adaptive transmissivity restoration image defogging method based on multi-scale window | |
| CN115170437A (en) | Fire scene low-quality image recovery method for rescue robot | |
| Abbaspour et al. | A new fast method for foggy image enhancement | |
| CN115619662A (en) | Image defogging method based on dark channel prior |
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 |