CN109886975A - It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network - Google Patents
It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network Download PDFInfo
- Publication number
- CN109886975A CN109886975A CN201910121153.4A CN201910121153A CN109886975A CN 109886975 A CN109886975 A CN 109886975A CN 201910121153 A CN201910121153 A CN 201910121153A CN 109886975 A CN109886975 A CN 109886975A
- Authority
- CN
- China
- Prior art keywords
- image
- raindrop
- network
- noise
- pixel
- 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
Landscapes
- Image Processing (AREA)
Abstract
Raindrop method and system is gone based on the image optimization processing for generating confrontation network the invention discloses a kind of, comprising: A, the original clear image of acquisition obtain test sample collection；B, original clear image is carried out by image processing software plus raindrop is handled, obtain training sample set；C, the band raindrop image that training sample is concentrated is subjected to noise reduction process；D, the image after noise reduction process is split, the multiple images after being divided；E, production network is constructed, clear image original in the image and training sample set after segmentation is inputted in production network and is trained the optimal output model of acquisition, finally the image after segmentation is input in optimal output model and obtains clear image；F, will clearly multiple images be spliced according to original clear image；G, processing finally is optimized to the image of splicing.The image that the present invention uses goes raindrop method easy to operate, at low cost, goes raindrop effect good, so that the image generated is more clear intuitively.
Description
Technical field
The present invention relates to images to go raindrop technical field, specially a kind of based on the image optimization processing for generating confrontation network
Remove raindrop method and system.
Background technique
Under the conditions of rainy day, captured image and video are easy scattering and fogging action by raindrop so that image at
As fuzzy, On The Deterioration of Visibility Over significantly limits the performance of outdoor vision processing algorithm.The image for causing camera to obtain is bad
Change, so that picture contrast is low, poor visibility, quality degradation.
Currently, image rain removing algorithm can be mainly divided into three classes: the first kind is to be based on image enhancement, but increase based on image
Certain information characteristics of image can be lost by force.Second class is the image restoration based on physical model, the purpose of Image Restoration Algorithm
It is the clearly image naturally for obtaining that there is good visibility, while keeping good color restorability；Based on hazy condition
The worsening reason of lower image establishes the physical model of atmospheric scattering, it is necessary first to estimate physical parameter model, as atmosphere light is shone
Intensity and transmissivity (depth), it is then inverse to solve the physical model to obtain no rain figure picture, but the image based on physical model is multiple
Original place reason is limited in scope；Third class is the image rain removing algorithm based on deep learning, such as convolutional neural networks are applied to image
Remove rain.Existing image rain removing method, which can be realized, goes rain to handle image, but low efficiency, and goes influence diagram after rain
The resolution ratio of picture, therefore, it is necessary to improve.
Summary of the invention
Raindrop method is gone based on the image optimization processing for generating confrontation network the purpose of the present invention is to provide a kind of, with solution
Certainly the problems mentioned above in the background art.
To achieve the above object, the invention provides the following technical scheme: it is a kind of based on the image optimization for generating confrontation network
Raindrop method is gone in processing, comprising the following steps:
A, original clear image is acquired, test sample collection is obtained；
B, original clear image is carried out by image processing software plus raindrop is handled, obtain training sample set；
C, the band raindrop image that training sample is concentrated is subjected to noise reduction process；
D, the image after noise reduction process is split, the multiple images after being divided；
E, production network is constructed, clear image original in the image and training sample set after segmentation is inputted into production
It is trained in network and obtains optimal output model, finally the image after segmentation is input in optimal output model and is obtained clearly
Image；
F, will clearly multiple images be spliced according to original clear image；
G, processing finally is optimized to the image of splicing.
Preferably, the specific embodiment that optimal output model is obtained in the step E is as follows:
Ea, pass through residual error network struction generator network module and arbiter network model；
Eb, the image after noise reduction process and segmentation is inputted in generator network model, exports image；
Ec, clear image original in training sample set is inputted in arbiter network model, obtains training neural network mould
Type；
Ed, the image for finally exporting step Eb input in training neural network model, obtain optimal output model.
Preferably, the specific embodiment that image optimization is handled in the step G is as follows:
Ga, the pixel of the image of output is divided into several figure layers according to brightness value, the brightness of each figure layer is different；
Gb, the figure layer minimum for brightness and the maximum figure layer of brightness first individually carry out histogram equalization processing, then
Ambient noise is removed, noise removal is finally carried out；
Gc, the figure layer among minimum brightness and maximum brightness is first removed into noise, then removes ambient noise, it is most laggard
Column hisgram equalization processing；
Gd, finally will it is processed after All Layers merge into the enhanced image of piece image.
Preferably, noise removal is handled using image interpolation functional operation in the step Gb and Gc, function formula
Are as follows: C '=A*T+D* (1-T), wherein C ' indicates image pixel after output denoising；A indicates present image pixel to be processed；T table
Show logic balance variable, T value is 0 or 1, the T=0 when the image pixel has raindrop feature, when the image pixel does not have
T=1 when raindrop feature；D indicates the noise smoothing value of currently pending pixel, and the method for median filtering, i.e. handle is employed herein
The value of current pixel is replaced with the intermediate value of point value each in neighborhood, to eliminate isolated noise spot.
Raindrop system is gone based on the image optimization processing for generating confrontation network the present invention also provides a kind of, including such as lower die
Block:
Test sample collection obtains module and obtains test sample collection for acquiring original clear image；
Training sample set obtains module, adds raindrop processing for carrying out original clear image by image processing software,
Obtain training sample set；
Noise reduction process module, the band raindrop image for concentrating training sample carry out noise reduction process；
Image segmentation module, for the image after noise reduction process to be split, the multiple images after being divided；
Production network constructs module, for constructing production network, by the image and training sample concentration after segmentation
It is trained in original clear image input production network and obtains optimal output model, be finally input to the image after segmentation
Clear image is obtained in optimal output model；
Image mosaic module, for will clearly multiple images be spliced according to original clear image；
Optimization processing module, for optimizing processing to the image of splicing.
Further, the specific embodiment that optimal output model is obtained in the production network building module is as follows,
Ea, pass through residual error network struction generator network model and arbiter network model；
Eb, the image after noise reduction process and segmentation is inputted in generator network model, exports image；
Ec, clear image original in training sample set is inputted in arbiter network model, obtains training neural network mould
Type；
Ed, the image for finally exporting step Eb input in training neural network model, obtain optimal output model.
Further, the specific embodiment that image optimization is handled in the optimization processing module is as follows,
Ga, the pixel of the image of splicing is divided into several figure layers according to brightness value, the brightness of each figure layer is different；
Gb, the figure layer minimum for brightness and the maximum figure layer of brightness first individually carry out histogram equalization processing, then
Ambient noise is removed, noise removal is finally carried out；
Gc, the figure layer among minimum brightness and maximum brightness is first removed into noise, then removes ambient noise, it is most laggard
Column hisgram equalization processing；
Gd, finally will it is processed after All Layers merge into the enhanced image of piece image.
Further, noise removal is handled using image interpolation functional operation in the step Gb and Gc, and function is public
Formula are as follows: C '=A*T+D* (1-T), wherein C ' indicates image pixel after output denoising；A indicates present image pixel to be processed；T
Indicate logic balance variable, T value is 0 or 1, the T=0 when the image pixel has raindrop feature, when the image pixel does not have
T=1 when having raindrop feature；D indicates the noise smoothing value of currently pending pixel, the method for median filtering is employed herein, i.e.,
The value of current pixel is replaced with the intermediate value of point value each in neighborhood, to eliminate isolated noise spot.
Compared with prior art, the beneficial effects of the present invention are:
(1) image that the present invention uses goes raindrop method easy to operate, at low cost, goes raindrop effect good；Wherein, of the invention
It is middle using image is split, go raindrop processing, splicing, enhancing processing by the way of, enable to generate image it is more clear
It is clear intuitive.
(2) the building production network method that the present invention uses can be improved network training efficiency, can obtain optimal figure
Picture.
(3) optimization method that the present invention uses reduces the global luminance difference of image, enhances picture contrast, effectively
Inhibit noise, further improve the clarity of image.
Detailed description of the invention
Fig. 1 is the method for the present invention overall flow figure；
Fig. 2 is image optimization process flow diagram in the method for the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
The present invention provides a kind of technical solution referring to FIG. 1-2: a kind of based on the image optimization processing for generating confrontation network
Go raindrop method, comprising the following steps:
A, original clear image is acquired, test sample collection is obtained；
B, original clear image is carried out by image processing software plus rain is handled, obtain training sample set；
C, the band raindrop image of acquisition is subjected to noise reduction process；
D, the image after noise reduction process is split, the multiple images after being divided；
E, production network is constructed, the image after segmentation is inputted in production network, and export clear image；
F, will clearly multiple images be spliced according to original clear image；
G, processing finally is optimized to the image of splicing.
In the present invention, it is as follows that optimal output model method is obtained in step E:
Ea, pass through residual error network struction generator network module and arbiter network model；
In Eb, the training sample input generator network model obtained after adding original clear image rain to handle, output figure
Picture；
Ec, training sample set is inputted in arbiter network model, obtains training neural network model；
Ed, the image for finally exporting step b input in training neural network model, obtain optimal output model.
The building production network method that the present invention uses can be improved network training efficiency, can obtain optimum image.
In the present invention, image optimization processing method is as follows in step G:
Ga, the pixel of the image of output being divided into several figure layers according to brightness value, the brightness of each figure layer is different, and
Each figure layer is pressed into brightness value, is arranged from low to high, and the boundary of the image in each figure layer is all by closed curve structure
At；
Gb, the figure layer minimum for brightness and the maximum figure layer of brightness first individually carry out histogram equalization processing, then
Ambient noise is removed, noise removal is finally carried out；
Gc, the figure layer among minimum brightness and maximum brightness is first removed into noise, then removes ambient noise, it is most laggard
Column hisgram equalization processing；
Gd, finally will it is processed after All Layers merge into the enhanced image of piece image.
Wherein, noise removal is handled using image interpolation functional operation, function formula are as follows: C '=A*T+D* (1-T),
Wherein, C ' indicates that image pixel after output denoising, A indicate present image pixel to be processed, and T indicates that logic balance variable, D indicate
The noise smoothing value of currently pending pixel.The optimization method that the present invention uses reduces the global luminance difference of image, enhancing
Picture contrast, effectively inhibits noise, further improves the clarity of image.
In conclusion the image that the present invention uses goes raindrop method easy to operate, and it is at low cost, go raindrop effect good；Wherein,
In the present invention by the way of being split image, going raindrop processing, splicing, enhancing processing, the image generated is enabled to
It is more clear.
The embodiment of the present invention is also provided a kind of handled based on the image optimization for generating confrontation network and goes raindrop system, including such as
Lower module:
Test sample collection obtains module and obtains test sample collection for acquiring original clear image；
Training sample set obtains module, adds raindrop processing for carrying out original clear image by image processing software,
Obtain training sample set；
Noise reduction process module, the band raindrop image for concentrating training sample carry out noise reduction process；
Image segmentation module, for the image after noise reduction process to be split, the multiple images after being divided；
Production network constructs module, for constructing production network, by the image and training sample concentration after segmentation
It is trained in original clear image input production network and obtains optimal output model, be finally input to the image after segmentation
Clear image is obtained in optimal output model；
Image mosaic module, for will clearly multiple images be spliced according to original clear image；
Optimization processing module, for optimizing processing to the image of splicing.
Further, the specific embodiment that optimal output model is obtained in the production network building module is as follows,
Ea, pass through residual error network struction generator network model and arbiter network model；
Eb, the image after noise reduction process and segmentation is inputted in generator network model, exports image；
Ec, clear image original in training sample set is inputted in arbiter network model, obtains training neural network mould
Type；
Ed, the image for finally exporting step Eb input in training neural network model, obtain optimal output model.
Further, the specific embodiment that image optimization is handled in the optimization processing module is as follows,
Ga, the pixel of the image of splicing is divided into several figure layers according to brightness value, the brightness of each figure layer is different；
Gb, the figure layer minimum for brightness and the maximum figure layer of brightness first individually carry out histogram equalization processing, then
Ambient noise is removed, noise removal is finally carried out；
Gc, the figure layer among minimum brightness and maximum brightness is first removed into noise, then removes ambient noise, it is most laggard
Column hisgram equalization processing；
Gd, finally will it is processed after All Layers merge into the enhanced image of piece image.
Further, noise removal is handled using image interpolation functional operation in the step Gb and Gc, and function is public
Formula are as follows: C '=A*T+D* (1-T), wherein C ' indicates image pixel after output denoising；A indicates present image pixel to be processed；T
Indicate logic balance variable, T value is 0 or 1, the T=0 when the image pixel has raindrop feature, when the image pixel does not have
T=1 when having raindrop feature；D indicates the noise smoothing value of currently pending pixel, the method for median filtering is employed herein, i.e.,
The value of current pixel is replaced with the intermediate value of point value each in neighborhood, to eliminate isolated noise spot.
It although an embodiment of the present invention has been shown and described, for the ordinary skill in the art, can be with
A variety of variations, modification, replacement can be carried out to these embodiments without departing from the principles and spirit of the present invention by understanding
And modification, the scope of the present invention is defined by the appended.
Claims (8)
1. a kind of go raindrop method based on the image optimization processing for generating confrontation network, which comprises the steps of:
A, original clear image is acquired, test sample collection is obtained；
B, original clear image is carried out by image processing software plus raindrop is handled, obtain training sample set；
C, the band raindrop image that training sample is concentrated is subjected to noise reduction process；
D, the image after noise reduction process is split, the multiple images after being divided；
E, production network is constructed, clear image original in the image and training sample set after segmentation is inputted into production network
In be trained and obtain optimal output model, finally the image after segmentation is input in optimal output model and is clearly schemed
Picture；
F, will clearly multiple images be spliced according to original clear image；
G, processing finally is optimized to the image of splicing.
2. a kind of handled based on the image optimization for generating confrontation network as described in claim 1 goes raindrop method, feature exists
In: the specific embodiment that optimal output model is obtained in the step E is as follows,
Ea, pass through residual error network struction generator network model and arbiter network model；
Eb, the image after noise reduction process and segmentation is inputted in generator network model, exports image；
Ec, clear image original in training sample set is inputted in arbiter network model, obtains training neural network model；
Ed, the image for finally exporting step Eb input in training neural network model, obtain optimal output model.
3. a kind of handled based on the image optimization for generating confrontation network as claimed in claim 1 or 2 goes raindrop method, feature
Be: the specific embodiment that image optimization is handled in the step G is as follows,
Ga, the pixel of the image of splicing is divided into several figure layers according to brightness value, the brightness of each figure layer is different；
Gb, the figure layer minimum for brightness and the maximum figure layer of brightness first individually carry out histogram equalization processing, then remove
Ambient noise finally carries out noise removal；
Gc, the figure layer among minimum brightness and maximum brightness is first removed into noise, then removes ambient noise, finally carried out straight
Square figure equalization processing；
Gd, finally will it is processed after All Layers merge into the enhanced image of piece image.
4. a kind of handled based on the image optimization for generating confrontation network as claimed in claim 3 goes raindrop method, feature exists
In: noise removal is handled using image interpolation functional operation in the step Gb and Gc, function formula are as follows: C '=A*T+D*
(1-T), wherein C ' indicates image pixel after output denoising；A indicates present image pixel to be processed；T indicates that logic balance becomes
Amount, T value are 0 or 1, the T=0 when the image pixel has raindrop feature, the T=when the image pixel does not have raindrop feature
1；D indicates the noise smoothing value of currently pending pixel, and the method for median filtering is employed herein, i.e., the value of current pixel is used
The intermediate value of each point value replaces in neighborhood, to eliminate isolated noise spot.
5. a kind of go raindrop system based on the image optimization processing for generating confrontation network, which is characterized in that including following module:
Test sample collection obtains module and obtains test sample collection for acquiring original clear image；
Training sample set obtains module, adds raindrop processing for carrying out original clear image by image processing software, obtains
Training sample set；
Noise reduction process module, the band raindrop image for concentrating training sample carry out noise reduction process；
Image segmentation module, for the image after noise reduction process to be split, the multiple images after being divided；
Production network constructs module will be original in the image and training sample set after segmentation for constructing production network
It is trained in clear image input production network and obtains optimal output model, be finally input to the image after segmentation optimal
Clear image is obtained in output model；
Image mosaic module, for will clearly multiple images be spliced according to original clear image；
Optimization processing module, for optimizing processing to the image of splicing.
6. a kind of handled based on the image optimization for generating confrontation network as claimed in claim 5 goes raindrop system, feature exists
In: the specific embodiment that optimal output model is obtained in the production network building module is as follows,
Ea, pass through residual error network struction generator network model and arbiter network model；
Eb, the image after noise reduction process and segmentation is inputted in generator network model, exports image；
Ec, clear image original in training sample set is inputted in arbiter network model, obtains training neural network model；
Ed, the image for finally exporting step Eb input in training neural network model, obtain optimal output model.
7. as a kind of handled based on the image optimization for generating confrontation network described in claim 5 or 6 goes raindrop system, feature
Be: the specific embodiment that image optimization is handled in the optimization processing module is as follows,
Ga, the pixel of the image of splicing is divided into several figure layers according to brightness value, the brightness of each figure layer is different；
Gb, the figure layer minimum for brightness and the maximum figure layer of brightness first individually carry out histogram equalization processing, then remove
Ambient noise finally carries out noise removal；
Gc, the figure layer among minimum brightness and maximum brightness is first removed into noise, then removes ambient noise, finally carried out straight
Square figure equalization processing；
Gd, finally will it is processed after All Layers merge into the enhanced image of piece image.
8. a kind of handled based on the image optimization for generating confrontation network as claimed in claim 7 goes raindrop system, feature exists
In: noise removal is handled using image interpolation functional operation in the step Gb and Gc, function formula are as follows: C '=A*T+D*
(1-T), wherein C ' indicates image pixel after output denoising；A indicates present image pixel to be processed；T indicates that logic balance becomes
Amount, T value are 0 or 1, the T=0 when the image pixel has raindrop feature, the T=when the image pixel does not have raindrop feature
1；D indicates the noise smoothing value of currently pending pixel, and the method for median filtering is employed herein, i.e., the value of current pixel is used
The intermediate value of each point value replaces in neighborhood, to eliminate isolated noise spot.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910121153.4A CN109886975A (en) | 2019-02-19 | 2019-02-19 | It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910121153.4A CN109886975A (en) | 2019-02-19 | 2019-02-19 | It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109886975A true CN109886975A (en) | 2019-06-14 |
Family
ID=66928417
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910121153.4A Pending CN109886975A (en) | 2019-02-19 | 2019-02-19 | It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109886975A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110267029A (en) * | 2019-07-22 | 2019-09-20 | 广州铭维软件有限公司 | A kind of long-range holographic personage's display technology based on AR glasses |
| CN110414362A (en) * | 2019-07-02 | 2019-11-05 | 安徽继远软件有限公司 | Electric power image data augmentation method based on production confrontation network |
| CN111738932A (en) * | 2020-05-13 | 2020-10-02 | 合肥师范学院 | Automatic rain removing method for photographed image of vehicle-mounted camera |
| CN112037187A (en) * | 2020-08-24 | 2020-12-04 | 宁波市眼科医院 | Intelligent optimization system for fundus low-quality pictures |
| CN112381728A (en) * | 2020-11-04 | 2021-02-19 | 天水师范学院 | Hyperspectral imaging signal processing system and method |
| CN114092492A (en) * | 2021-11-22 | 2022-02-25 | 广西师范大学 | High-definition color image segmentation algorithm |
| US11948279B2 (en) | 2020-11-23 | 2024-04-02 | Samsung Electronics Co., Ltd. | Method and device for joint denoising and demosaicing using neural network |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102456221A (en) * | 2010-10-25 | 2012-05-16 | 新奥特（北京）视频技术有限公司 | Method for rapidly eliminating image noise |
| CN104318542A (en) * | 2014-11-20 | 2015-01-28 | 上海华力创通半导体有限公司 | Image enhancement processing algorithm |
| CN106971378A (en) * | 2016-08-23 | 2017-07-21 | 上海海洋大学 | A kind of removing rain based on single image method based on depth denoising self-encoding encoder |
| CN108230278A (en) * | 2018-02-24 | 2018-06-29 | 中山大学 | A kind of image based on generation confrontation network goes raindrop method |
| CN108986044A (en) * | 2018-06-28 | 2018-12-11 | 广东工业大学 | A kind of image removes misty rain method, apparatus, equipment and storage medium |
-
2019
- 2019-02-19 CN CN201910121153.4A patent/CN109886975A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102456221A (en) * | 2010-10-25 | 2012-05-16 | 新奥特（北京）视频技术有限公司 | Method for rapidly eliminating image noise |
| CN104318542A (en) * | 2014-11-20 | 2015-01-28 | 上海华力创通半导体有限公司 | Image enhancement processing algorithm |
| CN106971378A (en) * | 2016-08-23 | 2017-07-21 | 上海海洋大学 | A kind of removing rain based on single image method based on depth denoising self-encoding encoder |
| CN108230278A (en) * | 2018-02-24 | 2018-06-29 | 中山大学 | A kind of image based on generation confrontation network goes raindrop method |
| CN108986044A (en) * | 2018-06-28 | 2018-12-11 | 广东工业大学 | A kind of image removes misty rain method, apparatus, equipment and storage medium |
Non-Patent Citations (2)
| Title |
|---|
| PHILLIP ISOLA等: "Image-to-Image Translation with Conditional Adversarial Networks", 《HTTPS://ARXIV.ORG/ABS/1611.07004》 * |
| RUI QIAN等: "Attentive Generative Adversarial Network for Raindrop Removal from a Single Image", 《HTTPS://ARXIV.ORG/ABS/1711.10098》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110414362A (en) * | 2019-07-02 | 2019-11-05 | 安徽继远软件有限公司 | Electric power image data augmentation method based on production confrontation network |
| CN110267029A (en) * | 2019-07-22 | 2019-09-20 | 广州铭维软件有限公司 | A kind of long-range holographic personage's display technology based on AR glasses |
| CN111738932A (en) * | 2020-05-13 | 2020-10-02 | 合肥师范学院 | Automatic rain removing method for photographed image of vehicle-mounted camera |
| CN112037187A (en) * | 2020-08-24 | 2020-12-04 | 宁波市眼科医院 | Intelligent optimization system for fundus low-quality pictures |
| CN112037187B (en) * | 2020-08-24 | 2024-03-26 | 宁波市眼科医院 | Intelligent optimization system for fundus low-quality pictures |
| CN112381728A (en) * | 2020-11-04 | 2021-02-19 | 天水师范学院 | Hyperspectral imaging signal processing system and method |
| US11948279B2 (en) | 2020-11-23 | 2024-04-02 | Samsung Electronics Co., Ltd. | Method and device for joint denoising and demosaicing using neural network |
| CN114092492A (en) * | 2021-11-22 | 2022-02-25 | 广西师范大学 | High-definition color image segmentation algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109886975A (en) | It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network | |
| CN110148095B (en) | Underwater image enhancement method and enhancement device | |
| CN106157267B (en) | Image defogging transmissivity optimization method based on dark channel prior | |
| Zhang et al. | A naturalness preserved fast dehazing algorithm using HSV color space | |
| CN103327220B (en) | With green channel for the denoising method guided on low-light (level) Bayer image | |
| CN108876743A (en) | A kind of image rapid defogging method, system, terminal and storage medium | |
| CN106530257A (en) | Remote sensing image de-fogging method based on dark channel prior model | |
| CN110148093B (en) | Image defogging improvement method based on dark channel prior | |
| CN102750674A (en) | Video image defogging method based on self-adapting allowance | |
| KR101426298B1 (en) | apparatus and method for compensating image for enhancing fog removing efficiency | |
| CN101626454B (en) | Method for intensifying video visibility | |
| CN113284064B (en) | Cross-scale context low-illumination image enhancement method based on attention mechanism | |
| CN108154492B (en) | A kind of image based on non-local mean filtering goes haze method | |
| WO2024060576A1 (en) | Image dehazing method based on dark channel prior | |
| CN107862672B (en) | Image defogging method and device | |
| CN111127340B (en) | Image defogging method | |
| CN110827225A (en) | Non-uniform illumination underwater image enhancement method based on double exposure frame | |
| CN111325688B (en) | Unmanned aerial vehicle image defogging method for optimizing atmosphere light by fusion morphology clustering | |
| CN107067375A (en) | A kind of image defogging method based on dark channel prior and marginal information | |
| CN111161167A (en) | Single image defogging method based on middle channel compensation and self-adaptive atmospheric light estimation | |
| CN112053298B (en) | Image defogging method | |
| CN110136079A (en) | Image defogging method based on scene depth segmentation | |
| CN114693548B (en) | Dark channel defogging method based on bright area detection | |
| CN110111268B (en) | Single image rain removing method and device based on dark channel and fuzzy width learning | |
| CN115760640A (en) | Coal mine low-illumination image enhancement method based on noise-containing Retinex model |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190614 |