CN111986119A - Dark channel image brightness value interference filtering method and sea fog image sea fog removing method - Google Patents
Dark channel image brightness value interference filtering method and sea fog image sea fog removing method Download PDFInfo
- Publication number
- CN111986119A CN111986119A CN202010906506.4A CN202010906506A CN111986119A CN 111986119 A CN111986119 A CN 111986119A CN 202010906506 A CN202010906506 A CN 202010906506A CN 111986119 A CN111986119 A CN 111986119A
- Authority
- CN
- China
- Prior art keywords
- dark channel
- straight line
- value
- brightness value
- final
- 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
- 238000000034 method Methods 0.000 title claims abstract description 25
- 238000001914 filtration Methods 0.000 title claims abstract description 10
- 239000002245 particle Substances 0.000 claims description 13
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 4
- 230000003287 optical effect Effects 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 2
- 230000000694 effects Effects 0.000 abstract description 4
- 238000012545 processing Methods 0.000 abstract description 4
- 238000005070 sampling Methods 0.000 abstract description 2
- 239000013535 sea water Substances 0.000 description 6
- 238000010586 diagram Methods 0.000 description 3
- 230000002238 attenuated effect Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- 238000002834 transmittance Methods 0.000 description 2
- 210000000746 body region Anatomy 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
Images
Classifications
-
- G06T5/73—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G06T5/70—
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The invention provides a dark channel image brightness value interference filtering method and a sea fog image sea fog removing method, and relates to the technical field of image processing, wherein the dark channel image brightness value interference filtering method adopts a random sampling consistency algorithm and a least square fitting algorithm to fit a brightness curve of a line dark channel image to obtain a dark pixel brightness value after interference removal; and the sea fog of the sea fog image is removed based on the dark pixel brightness value after the interference is removed, so that the defogging effect is improved.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a dark channel image brightness value interference filtering method and a sea fog image sea fog removing method.
Background
With the continuous development of unmanned aerial vehicle technology, the unmanned aerial vehicle has wider and wider application, and starts to engage in various patrol tasks including maritime patrol. When the unmanned aerial vehicle flies on the sea to execute patrol tasks, the unmanned aerial vehicle is easily interfered by sea fog, and therefore, the influence of the sea fog on the images shot by the unmanned aerial vehicle needs to be reduced by an image processing means.
An invention patent CN104933680B issued by the national intellectual property office in 2017, 10 months and 31 days discloses a method for removing image fog. In the scheme, the atmospheric brightness value is estimated according to the highest 0.01% of the brightness value in the image, and when no sky region (such as fig. 1) exists in the image, the estimated value is likely to be from a wave or a ship body region in fig. 1, which causes a deviation of the processing result from the actual situation. In addition, the sea water is not pure dark pixels, generally biased to blue, the R channel value is the lowest, but not close to 0, and the R channel value of the sea water pixels varies between 50 and 150 through a large number of real sea water image statistics. If the default sea water is a dark pixel, the sea water in the sea scene image after defogging is obviously bluish, which gives a sense of unreality, as shown in fig. 2.
Disclosure of Invention
Aiming at the technical problems, the invention provides a dark channel image brightness value interference filtering method and a sea fog image sea fog removing method.
The invention discloses a dark channel image brightness value interference filtering method, which comprises the following steps:
1. recording each line of brightness data in the dark channel image according to a point set, wherein each data point P in each line of imageiIncluding xi、diTwo dimensions, where xiRepresenting the horizontal coordinate of the data point in the line image, diA brightness value representing the data point;
2. randomly selecting partial data points in a certain row to obtain an initial consistent set S0;
3. performing straight line fitting on the brightness value of the initial consistent set S0 to obtain a straight line L0;
4. setting a distance threshold d _ th, screening data points in the initial consistent set S0, if the distance from a data point to a straight line L0 is greater than d _ th, discarding the point, removing all data points with the distance to the straight line L0 greater than d _ th to obtain a consistent set S1, and then performing straight line fitting on S1 to obtain L1;
5. reducing the distance threshold d _ th, repeating the step 4, continuously optimizing the consistent set until the consistent set is not changed any more, recording the optimal consistent set at the moment as S _ best, and recording a corresponding fitting straight line as L _ best;
6. calculating the distance from all data points in the line to a straight line L _ best, judging whether the data points belong to the optimal consistent set, discarding the data points which do not belong to the optimal consistent set to obtain a final consistent set S _ final and a fitting straight line L _ final thereof, and calculating the Ratio of the number of the data points in the final consistent set S _ final to the total number of the data points in the line to be Ratio _ final;
7. and repeating the step for sixteenth 1000 times, and selecting a fitting straight line L _ final corresponding to the final consistent set S _ final with the highest Ratio _ final value as a fitted brightness curve L of the line. Here, the N dark channel images correspond to N luminance curves L.
The invention also provides a sea fog removing method of the sea fog image, which comprises the following steps:
1. carrying out single-pixel dark channel extraction on the sea surface image, and taking the minimum value in the RGB value of each pixel as the dark channel value of the dark channel image;
2. fitting a brightness curve of each row of dark channel images by using a straight line through the dark channel image brightness value interference filtering method, and taking a brightness value corresponding to a middle point of the straight line obtained by fitting as a brightness value of the row;
3. linear modeling of transmittance ti＝ayi+ b, where tiDenotes the transmission coefficient, y, of the ith rowiRepresenting the y coordinate of the ith row in the dark channel image, and a and b are two linear parameters;
4. optical model according to fog I (x, y) ═ J (x, y) · t (x, y) + [1-t (x, y)]A (x, y) to obtain the ith row channel brightness value Ii＝J·ti+A(1-ti) Wherein J is the sea surface real brightness value in the dark channel image, and A is the atmospheric brightness value;
5. construction of energy values
Wherein N is the total number of dark channel image lines;
6. when the energy value E is calculated to be minimum by adopting a particle swarm evolution algorithm, corresponding values of four parameters J, A, a and b are used as optimal parameter values;
7. and calculating the real RGB value of each pixel according to the optimal parameter value.
The method adopts a random sampling consistency algorithm and a least square fitting algorithm to fit a brightness curve of a line dark channel image to obtain a dark pixel brightness value after interference is removed; and the sea fog of the sea fog image is removed based on the dark pixel brightness value after the interference is removed, so that the defogging effect is improved.
Drawings
FIG. 1 is an original unmanned aerial vehicle captured image;
FIG. 2 is a diagram illustrating the effect of the treatment after the default seawater of FIG. 1 is dark pixels;
FIG. 3 is the graph of FIG. 1 after single pixel dark channel extraction;
FIG. 4(a) is a graph of the luminance value of a dark channel image with an ordinate of 460;
FIG. 4(b) is a graph illustrating the luminance value curve of the dark channel image with ordinate 503;
FIG. 4(c) is a graph of the luminance value of the row-dark channel image with ordinate 611;
FIG. 4(d) is a graph of the luminance value of a dark channel image plotted on the ordinate 848;
FIG. 5 is a diagram of the fitted luminance curve L of FIG. 4 (a);
FIG. 6 is a diagram illustrating a curve in which luminance values of rows of a dark channel image are arranged in descending order;
FIG. 7 is a flow chart of a particle swarm evolution algorithm.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments. The embodiments of the present invention have been presented for purposes of illustration and description, and are not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art. The embodiment was chosen and described in order to best explain the principles of the invention and the practical application, and to enable others of ordinary skill in the art to understand the invention for various embodiments with various modifications as are suited to the particular use contemplated.
For ease of presentation, the underlying concepts will now be described in advance as follows:
optical model of fog
There are two parts of attenuation in the brightness values of the images taken by the camera: 1. atmospheric brightness is attenuated by foggy air; 2. the reflected light of the scene object is attenuated by the fog, so that I (x, y) ═ J (x, y) · t (x, y) + [1-t (x, y) ] · a (x, y), where J (x, y) is the image true luminance value, t (x, y) is the transmission coefficient, and a (x, y) is the atmospheric luminance value.
Second, particle swarm evolution algorithm
1. Individual/particle: each individual corresponds to a group of solutions, and in the problem to be solved by the invention, each individual comprises four parameters of J, A, a and b, and each individual corresponds to a fitness.
2. Group: the population comprises a plurality of individuals, e.g., 50, 100, etc., the specific number depending on the amount of computing resources.
3. Fitness is as follows: after an optimization problem is defined, the fitness is related to the target value of the problem, and the closer to the target value, the higher the fitness.
4. Generation: the evolutionary computing method is used for searching the optimal solution of the optimization problem, and needs to perform a plurality of iterations, wherein each iteration corresponds to a generation. The same individual in different generations has different parameter values.
5. Inheritance: when the individual parameters are iterated, the parameter values of the next generation can be greatly referred to the parameter values of the current generation, and the genetic rules of organisms are similar.
6. Mutation: when the individual parameters are iterated, the parameter values of the next generation can introduce new variation, and have a certain difference with the parameter values of the current generation, which is similar to the variation rule of organisms.
7. Individual history optimal record: and in the process of multiple iterations, the parameter value corresponding to the highest fitness historically presented by the current individual.
8. Optimal record of group history: and in the process of multiple iterations, all individuals in the group historically show parameter values corresponding to the highest fitness.
9. Speed: the difference of the current individual relative to each parameter of the previous generation, and the calculation formula of the current particle velocity in iteration is vi+1＝w×vi+c1r1×(pbesti-xi)+c2r2(gbesti-xi) The method is obtained by adding the following three parts:
first generation speed w × vi: velocity v of the current generationi+1Reference to the velocity v of the previous generationiW is a weight value, set empirically, e.g., w is 0.8, with different weight values determining how well the particles change in the iterative process follows the historical change law.
Optimum record c of individual history1r1×(pbesti-xi): the particles refer to their historical optimal positions when changing. pbestiMeans the individual optimum position, i is the number of the particle, xiIs the position of the current particle in the feature space, i.e. the parameter value, c1Is a constant value r1Is a random number between 0 and 1, randomly generated at each iteration.
This means, in part, that individuals try to move closer to their historical optimal location as they change, since the perimeter of the historical optimal location means that better fitness is possible.
History optimal record c of the group2r2(gbesti-xi): the particles, when changing, reference the historically optimal location of all individuals of the entire population. gbestiIs the optimal position of the population, c2Is a constant value r2Is a random number between 0 and 1, randomly generated at each iteration.
This means that individuals try to get closer to the group history optimal location when they change, because the perimeter of the group history optimal location means that better fitness is possible.
10. Position: the parameter value of the current individual. After obtaining the individual speed, the next generation position x can be obtained by only adding the speed to the current positioni+1＝xi+vi+1。
The method for removing sea fog from sea fog images disclosed by the invention comprises the following steps:
1. and (3) carrying out single-pixel dark channel extraction on the sea surface image, and taking the minimum value in the RGB values of each pixel as the dark channel value of the dark channel image to obtain the image shown in the figure 3.
The dark channel values in each row of dark channel image of fig. 3 are not completely consistent, but are disturbed by various factors, such as brightness variations caused by wave fluctuation, white waves, pixels with different brightness on the ship, and the like.
Fig. 4(a) -4(d) are four rows of dark channel image luminance value curves whose ordinate in fig. 3 (the ordinate of the uppermost row is 1, and which increase from top to bottom) is 460, 503, 611, 848, respectively. It is easy to see that the previous three rows of dark channel images are greatly changed in curve because of the interference of ships or wave flowers, and the brightness value of the real dark pixel of the row is calculated to have larger interference; the fourth row is almost sea level, the impurity interference is less, and the brightness curve is much flatter.
Therefore, the interference factors must be filtered to obtain the real dark channel brightness value of each row of dark channel images, and the overall defogging effect is improved.
2. Fitting a brightness curve of each row of dark channel images by using a straight line, and taking a brightness value corresponding to a middle point of the straight line obtained by fitting as a brightness value of the row, wherein the method specifically comprises the following steps:
recording is performed according to a set of points for each row of luminance data in a dark channel image, each data point P in each row of imageiIncluding xi、diTwo dimensions, where xiRepresenting the horizontal coordinate of the data point in the line image, diRepresenting the brightness value of the data point.
Next, for a particular row, randomly selecting some (e.g., half) of the data points within the row results in an initial consistent set S0.
And performing straight line fitting on the brightness value of the initial consistent set S0 to obtain a straight line L0.
Fourth, a distance threshold d _ th is set, data points in the initial consistent set S0 are screened, if the distance from the data points to the straight line L0 is larger than d _ th, the data points are discarded, all data points with the distance to the straight line L0 larger than d _ th are removed, a consistent set S1 is obtained, and then straight line fitting is conducted on S1, and L1 is obtained.
Fifthly, reducing the distance threshold d _ th, repeating the step (4), continuously optimizing the consistent set until the consistent set is not changed any more, recording the best consistent set at the moment as S _ best, and recording a corresponding fitting straight line as L _ best.
Sixthly, calculating the distance from all data points in the line to a straight line L _ best, judging whether the data points belong to the optimal consistent set, discarding the data points which do not belong to the optimal consistent set to obtain a final consistent set S _ final and a fitting straight line L _ final thereof, and calculating the Ratio of the number of the data points in the final consistent set S _ final to the total number of the data points in the line to be Ratio _ final.
And repeating the steps from sixteenth to sixteenth 1000 times, and selecting a fitting straight line L _ final corresponding to the final consistent set S _ final with the highest Ratio _ final value as a fitted brightness curve L of the line. The bottom straight line in fig. 5 is the luminance curve L after fitting of fig. 4 (a). Here, the N dark channel images correspond to N luminance curves L.
4. The luminance values of the rows of the dark channel image are arranged in the order of the ordinate from top to bottom, and a curve is drawn, as shown in fig. 6. Since the sea surface color is not perfectly uniform, the curve in fig. 6 is only approximately a straight line, approximately close to a linear gradient, but it is sufficient to use a linear model to simulate the transmission law of fog on the sea surface.
Thus, the transmittance is linearly modeled ti＝ayi+ b, where tiDenotes the transmission coefficient, y, of the ith rowiWhich represents the y-coordinate of the ith row in the dark channel image, a, b are two linear parameters.
5. Optical model according to fog I (x, y) ═ J (x, y) · t (x, y) + [1-t (x, y)]A (x, y) to obtain the ith row channel brightness value Ii＝J·ti+A(1-ti) Wherein J is the sea surface real brightness value in the dark channel image, and A is the atmospheric brightness value.
6. Construction of energy values
Where N is the total number of dark channel image lines.
7. When the energy value E is calculated to be minimum by adopting the particle swarm evolution algorithm, corresponding optimal parameter values of the four parameters J, A, a and b are shown in a working flow chart of the particle swarm algorithm in fig. 7, and the iteration process lasts for a plurality of generations until the energy value E reaches the satisfaction degree.
8. And calculating the real RGB value of each pixel according to the optimal parameter value.
Obtaining transmission coefficient t of each row by using parameters a and bi＝ayi+ b, according to y for a particular rowiIs a known number;
obtaining the real RGB value of each pixel position in the clear image after defogging
Wherein O isiR、OiG、OiBIs the original sea fog image RGB value.
It is to be understood that the described embodiments are merely a few embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by one of ordinary skill in the art and related arts based on the embodiments of the present invention without any creative effort, shall fall within the protection scope of the present invention. It is to be understood that the described embodiments are merely a few embodiments of the invention, and not all embodiments. All other embodiments, which can be derived by one of ordinary skill in the art and related arts based on the embodiments of the present invention without any creative effort, shall fall within the protection scope of the present invention.
Claims (2)
1. A dark channel image brightness value interference filtering method is characterized by comprising the following steps:
step A1, recording each line brightness data in the dark channel image according to a point set, each data point P in the line imageiIncluding xi、diTwo dimensions, where xiRepresenting the horizontal coordinate of the data point in the line image, diA brightness value representing the data point;
step A2, aiming at a specific row, randomly selecting partial data points in the row to obtain an initial consistent set S0;
step A3, performing straight line fitting on the brightness values of the initial consistent set S0 to obtain a straight line L0;
step A4, setting a distance threshold d _ th, screening data points in an initial consistent set S0, if the distance from a data point to a straight line L0 is greater than d _ th, discarding the point, removing all data points with the distance to the straight line L0 being greater than d _ th to obtain a consistent set S1, and then performing straight line fitting on S1 to obtain L1;
step A5, reducing the distance threshold d _ th, repeating the step A4, continuously optimizing a consistent set until the consistent set is not changed any more, recording the optimal consistent set at the moment as S _ best, and recording a corresponding fitting straight line as L _ best;
step A6, calculating the distance from all data points in the line to the straight line L _ best, judging whether the data points belong to the optimal consistent set, discarding the data points which do not belong to the optimal consistent set to obtain a final consistent set S _ final and a fitting straight line L _ final thereof, and calculating the Ratio of the number of the data points in the final consistent set S _ final to the total number of the data points in the line to be Ratio _ final;
and step A7, repeating the steps A2 to A6 for a plurality of times, and selecting the fitting straight line L _ final corresponding to the final consistent set S _ final with the highest Ratio _ final value as the fitted brightness curve L of the line.
2. A sea fog removing method for sea fog images is characterized by comprising the following steps:
step 1, carrying out single-pixel dark channel extraction on sea surface images, and taking the minimum value of RGB values of each pixel as a dark channel value of a dark channel image;
step 2, fitting a brightness curve of the dark channel image of each row by using a straight line through the dark channel image brightness value interference filtering method of claim 1, and taking the brightness value corresponding to the middle point of the fitted straight line as the brightness value of the row;
step 3, carrying out linear modeling t on the transmissivityi＝ayi+ b, where tiDenotes the transmission coefficient, y, of the ith rowiRepresenting the y coordinate of the ith row in the dark channel image, and a and b are two linear parameters;
step 4, according to the optical model of fog, I (x, y) ═ J (x, y) · t (x, y) + [1-t (x, y)]A (x, y) to obtain the ith row channel brightness value Ii＝J·ti+A(1-ti) Wherein J is the sea surface real brightness value in the dark channel image, and A is the atmospheric brightness value;
step 5, constructing an energy value
Wherein N is the total number of dark channel image lines;
step 6, when the energy value E is calculated to be minimum by adopting a particle swarm evolution algorithm, the corresponding values of the four parameters J, A, a and b are used as optimal parameter values;
and 7, calculating the real RGB value of each pixel according to the optimal parameter value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010906506.4A CN111986119B (en) | 2020-09-01 | 2020-09-01 | Interference filtering method for brightness value of dark channel image and sea fog removing method for sea fog image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010906506.4A CN111986119B (en) | 2020-09-01 | 2020-09-01 | Interference filtering method for brightness value of dark channel image and sea fog removing method for sea fog image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111986119A true CN111986119A (en) | 2020-11-24 |
| CN111986119B CN111986119B (en) | 2023-11-28 |
Family
ID=73447376
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010906506.4A Active CN111986119B (en) | 2020-09-01 | 2020-09-01 | Interference filtering method for brightness value of dark channel image and sea fog removing method for sea fog image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111986119B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105681774A (en) * | 2016-02-26 | 2016-06-15 | 深圳市金立通信设备有限公司 | Image processing method and terminal |
| KR101705536B1 (en) * | 2015-10-08 | 2017-02-10 | 목포해양대학교 산학협력단 | A fog removing method based on camera image |
| CN106548463A (en) * | 2016-10-28 | 2017-03-29 | 大连理工大学 | Based on dark and the sea fog image automatic defogging method and system of Retinex |
| CN108416815A (en) * | 2018-03-05 | 2018-08-17 | 国家安全生产监督管理总局通信信息中心 | Assay method, equipment and the computer readable storage medium of air light value |
| US20190114747A1 (en) * | 2016-04-07 | 2019-04-18 | Carmel Haifa University Economic Corporation Ltd. | Image dehazing and restoration |
| CN110135434A (en) * | 2018-11-13 | 2019-08-16 | 天津大学青岛海洋技术研究院 | Underwater picture increased quality algorithm based on color line model |
-
2020
- 2020-09-01 CN CN202010906506.4A patent/CN111986119B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101705536B1 (en) * | 2015-10-08 | 2017-02-10 | 목포해양대학교 산학협력단 | A fog removing method based on camera image |
| CN105681774A (en) * | 2016-02-26 | 2016-06-15 | 深圳市金立通信设备有限公司 | Image processing method and terminal |
| US20190114747A1 (en) * | 2016-04-07 | 2019-04-18 | Carmel Haifa University Economic Corporation Ltd. | Image dehazing and restoration |
| CN106548463A (en) * | 2016-10-28 | 2017-03-29 | 大连理工大学 | Based on dark and the sea fog image automatic defogging method and system of Retinex |
| CN108416815A (en) * | 2018-03-05 | 2018-08-17 | 国家安全生产监督管理总局通信信息中心 | Assay method, equipment and the computer readable storage medium of air light value |
| CN110135434A (en) * | 2018-11-13 | 2019-08-16 | 天津大学青岛海洋技术研究院 | Underwater picture increased quality algorithm based on color line model |
Non-Patent Citations (5)
| Title |
|---|
| 毕雪英;过洁;潘金贵;: "强光源影响下的单幅图像去雾", 常州大学学报(自然科学版), no. 01 * |
| 王飞;刘梦婷;刘雪芹;秦志亮;马本俊;郑毅;: "基于YOLOv3深度学习的海雾气象条件下海上船只实时检测", 海洋科学, no. 08 * |
| 郑思凡;王卫星;何占华;梁子裕;陈平平;: "双粒度光流流形学习的刮刷总成摆杆摆幅检测", 华南理工大学学报(自然科学版), no. 01 * |
| 陈钟荣;张炎;张瑶;: "基于自适应雾浓度系数的暗通道先验法能见度测量", 现代电子技术, no. 09 * |
| 韩静;于平平;王震洲;霍威;齐林;: "基于相机响应曲线拟合的道路能见度检测算法", 燕山大学学报, no. 03 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111986119B (en) | 2023-11-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112288658A (en) | Underwater image enhancement method based on multi-residual joint learning | |
| CN108564549B (en) | Image defogging method based on multi-scale dense connection network | |
| CN111292257B (en) | Retinex-based image enhancement method in scotopic vision environment | |
| CN109493300B (en) | Aerial image real-time defogging method based on FPGA (field programmable Gate array) convolutional neural network and unmanned aerial vehicle | |
| CN110097106A (en) | The low-light-level imaging algorithm and device of U-net network based on deep learning | |
| Mei et al. | Single image dehazing using dark channel fusion and haze density weight | |
| CN114972102A (en) | Underwater image enhancement method based on global variable contrast enhancement and local correction | |
| Cai et al. | Underwater image processing system for image enhancement and restoration | |
| CN111986119B (en) | Interference filtering method for brightness value of dark channel image and sea fog removing method for sea fog image | |
| Li et al. | Global color consistency correction for large-scale images in 3-D reconstruction | |
| CN109949239B (en) | Self-adaptive sharpening method suitable for multi-concentration multi-scene haze image | |
| CN110349113B (en) | Adaptive image defogging method based on dark primary color priori improvement | |
| CN116385293A (en) | Foggy-day self-adaptive target detection method based on convolutional neural network | |
| CN112950484B (en) | Method for removing color pollution of photographic image | |
| CN115393209A (en) | Bayesian Retinex deep sea image enhancement method based on multi-constraint prior | |
| CN111640082B (en) | Underwater image recovery method based on Gaussian mixture model and dark channel theory | |
| CN112907503B (en) | Penaeus vannamei Boone quality detection method based on self-adaptive convolutional neural network | |
| CN114663683A (en) | Underwater target detection method based on spatial feature self-supervision | |
| CN113191971A (en) | Unmanned aerial vehicle image defogging method based on YUV color space | |
| CN112686825A (en) | Ground dense fog scene gray image defogging method based on FPGA | |
| CN111489302A (en) | Maritime image enhancement method in fog environment | |
| CN115496694B (en) | Method for recovering and enhancing underwater image based on improved image forming model | |
| CN111640081B (en) | Underwater image recovery method based on optimization and dark channel | |
| CN115205668A (en) | Method and system for detecting netting attachment based on deep learning | |
| CN113781329B (en) | Fog removing method for remote sensing image |
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 |