CN108986044A - A kind of image removes misty rain method, apparatus, equipment and storage medium - Google Patents
A kind of image removes misty rain method, apparatus, equipment and storage medium Download PDFInfo
- Publication number
- CN108986044A CN108986044A CN201810688056.9A CN201810688056A CN108986044A CN 108986044 A CN108986044 A CN 108986044A CN 201810688056 A CN201810688056 A CN 201810688056A CN 108986044 A CN108986044 A CN 108986044A
- Authority
- CN
- China
- Prior art keywords
- image
- rain
- deep learning
- data set
- misty rain
- 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 title claims abstract description 57
- 238000003860 storage Methods 0.000 title claims abstract description 13
- 238000013135 deep learning Methods 0.000 claims abstract description 47
- 239000003595 mist Substances 0.000 claims abstract description 43
- 238000012549 training Methods 0.000 claims abstract description 38
- 238000012545 processing Methods 0.000 claims abstract description 17
- 230000000694 effects Effects 0.000 claims abstract description 16
- 230000006870 function Effects 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 7
- 230000004069 differentiation Effects 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 3
- 238000012986 modification Methods 0.000 claims description 3
- 241000208340 Araliaceae Species 0.000 claims description 2
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 claims description 2
- 235000003140 Panax quinquefolius Nutrition 0.000 claims description 2
- 235000008434 ginseng Nutrition 0.000 claims description 2
- 238000003707 image sharpening Methods 0.000 abstract description 5
- 235000013399 edible fruits Nutrition 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000002360 preparation method Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/77—Retouching; Inpainting; Scratch removal
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
This application discloses a kind of images to remove misty rain method, apparatus, equipment and storage medium, this method comprises: acquiring data set first;The data set includes clear figure, and clearly schemes matched rain figure and mist figure;Then collected data set training deep learning CGAN network is utilized;Misty rain is carried out finally by trained deep learning CGAN network handles processing image to handle.The ambiguity being likely to result in view of the generation of image in weather misty rain, the application directly learns many rain figures and mist figure using deep learning CGAN network, when the image to be processed no matter inputted is rain figure or mist figure, it can effectively identify, preferably utilize the feature of whole figure, lose image feature information not, original image is restored well by the correlation of image, reach good sharpening effect, with more reliability and accuracy, and most blurred picture can be handled, image sharpening is easier, be easier.
Description
Technical field
The present invention relates to technical field of image processing, go misty rain method, apparatus, equipment more particularly to a kind of image and deposit
Storage media.
Background technique
When taking pictures image in greasy weather or rainy day, it may be formed a little fuzzy and blocked when at figure, caused pair
The understanding of image is there may be difficulty, and at this moment the sharpening of image is with regard to necessary.
And the existing method for going rain or defogging about image, it is based only on the place carried out under a kind of fuzzy enviroment to image
Reason can only handle a kind of specified rain figure either mist figure, this will achieve the effect that or defogging for raining in unknown situation
It is difficult precognition, an image automatic identification can not be reached by a program goes rain or defogging.This is because existing
Method all only do the segmentation of image, the feature extraction of feature or mist only for rain, this whole benefit for image
With comparatively seldom, so processing can only achieve single effect.
Therefore, how automatic identification rain figure, mist figure and progress effectively remove misty rain, are that those skilled in the art are urgently to be resolved
Technical problem.
Summary of the invention
In view of this, the purpose of the present invention is to provide a kind of images to remove misty rain method, apparatus, equipment and storage medium,
More convenient, more acurrate by deep learning CGAN network image to be processed can be carried out containing mist and containing the processing of rain.It is specific
Scheme is as follows:
A kind of image goes misty rain method, comprising:
Acquire data set;The data set includes clear figure, clearly schemes matched rain figure and mist figure with described;
Utilize the collected data set training deep learning CGAN network;
Misty rain is carried out by the trained deep learning CGAN network handles processing image to handle.
Preferably, it is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, utilizes the collected data
Collect training deep learning CGAN network, specifically include:
Classify to the collected data set, is divided into training set and verifying collection;
The training set is input to deep learning CGAN network to be trained, automatically adjusts the deep learning CGAN net
The parameter of generator and arbiter in network;
The output of the deep learning CGAN network is observed in the training process as a result, judging whether by the verifying collection
It needs artificially to modify hyper parameter and initialization value.
Preferably, it is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, the training set is input to depth
Degree study CGAN network is trained, and automatically adjusts the parameter of generator and arbiter in the deep learning CGAN network, has
Body includes:
Building condition generates deep learning CGAN network;The deep learning CGAN network include generator and arbiter this
Two models;
Read the training set, and by the training set of reading and random noise Input matrix to the generator and life
At new images;
The new images that the generator generates are compared with the clear figure by the arbiter, are tied according to comparison
Fruit calculates loss function;
Penalty values are calculated according to the loss function, automatic training simultaneously continues to adjust the generator and the arbiter
Parameter, until the value relative minimum or epoch for making the loss function reach preset threshold.
Preferably, it is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, adjusts the generator and described
The parameter of arbiter, specifically includes:
The parameter for adjusting the generator makes the new images generated in the generator close to the clear figure, not by
Erroneous judgement is other;
The parameter for adjusting the arbiter keeps the differentiation probability of the clear figure in loss function bigger, the generator
The differentiation probability of the new images of interior generation is bigger.
Preferably, it is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, acquires data set, specifically include:
Collected clear figure is stored in the first file, will clearly scheme matched rain figure and mist figure deposit second with described
File;
Identical name is carried out to the clear figure, the rain figure and the mist figure that are mutually matched;
The image size being stored into first file and second file is adjusted to unified picture format,
As data set gathered in advance.
Preferably, it is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, clearly schemes matched rain with described
Figure is to carry out that rain effect is added to obtain to the clear figure;
With it is described clearly scheme matched mist figure be the clear figure is carried out plus fog effect obtain.
The embodiment of the invention also provides a kind of images to remove misty rain device, comprising:
Dataset acquisition module, for acquiring data set;The data set includes clear figure, with it is described clearly scheme it is matched
Rain figure and mist figure;
Network training module, for utilizing the collected data set training deep learning CGAN network;
Image processing module carries out rain for handling image by the trained deep learning CGAN network handles
Mist processing.
The embodiment of the invention also provides a kind of images to remove misty rain equipment, including processor and memory, wherein the place
Reason device is realized when executing the computer program saved in the memory as above-mentioned image provided in an embodiment of the present invention removes misty rain
Method.
The embodiment of the invention also provides a kind of computer readable storage mediums, for storing computer program, wherein institute
It states and is realized when computer program is executed by processor as above-mentioned image provided in an embodiment of the present invention goes misty rain method.
A kind of image provided by the present invention removes misty rain method, apparatus, equipment and storage medium, this method comprises: first
Acquire data set;The data set includes clear figure, clearly schemes matched rain figure and mist figure with described;Then collected number is utilized
Deep learning CGAN network is trained according to collecting;Rain is carried out finally by trained deep learning CGAN network handles processing image
Mist processing.
In view of the ambiguity that the generation of image is likely to result in weather misty rain, the present invention uses deep learning
CGAN network directly learns many rain figures and mist figure, can be effective when the image to be processed no matter inputted is rain figure or mist figure
Identification preferably utilizes the feature of whole figure, loses image feature information not, restored well by the correlation of image
Original image reaches good sharpening effect, reliability and accuracy is had more, and can handle most blurred picture, by image
Sharpening is easier, is easier.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the flow chart that image provided in an embodiment of the present invention goes misty rain method;
Fig. 2 is the structural schematic diagram that image provided in an embodiment of the present invention removes misty rain device.
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 image and goes misty rain method, as shown in Figure 1, comprising the following steps:
S101, acquisition data set;The data set includes clear figure, and clearly schemes matched rain figure and mist figure;
S102, collected data set training deep learning CGAN network is utilized;
S103, image is handled by trained deep learning CGAN network handles carry out misty rain and handle.
It is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, acquisition first includes clear figure, with clear figure
The data set of matched rain figure and mist figure;Then network is fought using collected data set training deep learning convolution production
(Convolutional Generative Adversarial Network, abbreviation CGAN);Finally by trained depth
Study CGAN network handles processing image carries out misty rain and handles.Directly learn many rain using deep learning CGAN network in this way
Figure and mist figure can be identified effectively when the image to be processed of input, which either contains rain figure, still contains mist figure, preferably using whole
The feature for opening figure, loses image feature information not, restores original image well by the correlation of image, reaches clear well
Clearization effect has more reliability and accuracy, and can handle most blurred picture, image sharpening is easier, more hold
Easily.
Further, due to the deep learning CGAN network of use, in training, to need to input two images having the same
Information, one containing rainy or mist, another Zhang Ze is clearly image, this is difficult to exist in reality scene, so before needing to pass through
The preparation of phase obtains relevant data set.Therefore, in the specific implementation, step S101 acquire data set, can specifically include with
Lower step:
Step 1: the original image of the clear image under good environment is found, as wait locate by finding online public data collection
Manage the clear figure of image;
Step 2: carrying out adding fog effect and adding for image without rain fog free images to searching using traditional images processing method
The processing of rain effect, i.e., with clearly scheme matched rain figure be clear figure is carried out plus rain effect obtain, with clearly scheme it is matched
Mist figure is to carry out that fog effect is added to obtain to clear figure;
Step 3: collected clear figure is stored in the first file, will with clearly scheme matched rain figure and mist figure is stored in
Second file;Identical name is carried out to the clear figure, rain figure and the mist figure that are mutually matched, guarantees clear figure, contain rain or figure containing mist
As reaching one-to-one correspondence, could accomplishing when establishing model in this way to the corresponding extraction of the feature of rain figure or mist figure and eliminating;
Step 4: carrying out picture reconstruction, the image size being stored into the first file and the second file is adjusted to unite
One picture format, such as the picture format of 512*512, as data set gathered in advance.
Further, in the specific implementation, it is gone in misty rain method in above-mentioned image provided in an embodiment of the present invention, depth
Practising CGAN network includes two models of generator and arbiter, and step S102 utilizes collected data set training deep learning
CGAN network, can specifically include following steps:
Firstly, classifying to collected data set, the data set of acquirement is divided into training set and verifying collects;
Later, training set is input to deep learning CGAN network to be trained, automatically adjusts deep learning CGAN network
The parameter of middle generator and arbiter needs training set to train the weighting parameter that deep learning can be modified independently;
Finally, observing the output of deep learning CGAN network in the training process by verifying collection as a result, judging whether to need
Very important person is modification hyper parameter and initialization value;If desired, then the modification to hyper parameter and initialization value is executed, if not needing,
Directly export network model.
For the parameter for automatically adjusting generator and arbiter in above-mentioned steps, it can specifically include following steps:
The first step, building condition generate deep learning CGAN network;Deep learning CGAN network includes generator and sentences
Other the two models of device;If G is generator, D is arbiter.Specifically, the related training parameter of CGAN network is set, for example,
The number of learning rate and epoch can adjust the training effect of entire CGAN network;
Second step is read training set (all kinds of image files in the first file and the second file), if the first file
Interior clear figure (clearly without misty rain original image) is y, in the second file with clearly to scheme matched rain figure or mist figure be x, make an uproar at random
Sound matrix is z;By in training set rain figure or mist figure x and random noise matrix z after be input to generator G, and generate new images
z*;
Third step, the new images z for generating generator G*It is compared with clear figure y by arbiter D, is tied according to comparison
Fruit calculates loss function:
Wherein, E represents the mean value of pictures in each bacth_size, and D (y | x) is represented using mist figure or rain figure as marking
Clear figure y is determined as to the probability of no misty rain image when label;D(G(z*| x)) represent by mist figure or rain figure will to be given birth to when label
At new images z*It is determined as the probability of no misty rain image.
4th step calculates penalty values according to loss function, automatic training and the ginseng for continuing to adjust generator and arbiter
Number;Second step is repeated always to the 4th step, until the value of loss function is relative minimum or epoch reaches preset threshold and is
Only, detailed process is as follows:
The parameter for adjusting generator G first, makes the new images z generated in generator G*It more can be (clear close to clear figure
Without misty rain original image), it is not misjudged not, i.e. G (z | x) is smaller;Then the parameter for adjusting arbiter D, it is clear in loss function to make
Scheme differentiation probability D (y | x) bigger, new images z of generation of y*Differentiation probability D (G (z*| x)) it is bigger;Thus constantly adjustment two
The parameter of a network, it is minimum to meet G (z | x) probability, D (y | x) and D (G (z*| x)) maximum probability, finally obtain loss function
Relative minimum MinMax:
5th step obtains one and removes misty rain network model G based on CGAN, image to be processed (can be had the figure of misty rain
Piece) as input, ultimately generate the high definition figure of no misty rain.
Based on the same inventive concept, the embodiment of the invention also provides a kind of images to remove misty rain device, since the image is gone
The principle that misty rain device solves the problems, such as goes misty rain method similar to a kind of aforementioned image, therefore the image goes the implementation of misty rain device
It may refer to the implementation that image goes misty rain method, overlaps will not be repeated.
In the specific implementation, image provided in an embodiment of the present invention removes misty rain device, as shown in Fig. 2, specifically including:
Dataset acquisition module 11, for acquiring data set;Data set includes clear figure, with clearly scheme matched rain figure and
Mist figure;
Network training module 12, for utilizing collected data set training deep learning CGAN network;
Image processing module 13 carries out misty rain for handling image by trained deep learning CGAN network handles
Processing.
It is gone in misty rain device in above-mentioned image provided in an embodiment of the present invention, the phase interaction of above three module can be passed through
With can handle rain figure and mist figure simultaneously, when the image to be processed no matter inputted is rain figure or mist figure, can effectively identify, fill
Divide the characteristic information using whole figure, lose image feature information not, original image is restored by the correlation of image well,
Reach good sharpening effect, reliability and accuracy is had more, and most blurred picture can be handled, by image sharpening
It is easier, be easier.
Corresponding contents disclosed in previous embodiment can be referred to about the more specifical course of work of above-mentioned modules,
This is no longer repeated.
Correspondingly, the embodiment of the invention also discloses a kind of images to remove misty rain equipment, including processor and memory;Its
In, realize that image disclosed in previous embodiment goes misty rain method when processor executes the computer program saved in memory.
It can be with reference to corresponding contents disclosed in previous embodiment, herein no longer about the more specifical process of the above method
It is repeated.
Further, the invention also discloses a kind of computer readable storage mediums, for storing computer program;It calculates
Machine program realizes that aforementioned disclosed image goes misty rain method when being executed by processor.
It can be with reference to corresponding contents disclosed in previous embodiment, herein no longer about the more specifical process of the above method
It is repeated.
Each embodiment in this specification is described in a progressive manner, the highlights of each of the examples are with it is other
The difference of embodiment, same or similar part may refer to each other between each embodiment.For being filled disclosed in embodiment
It sets, for equipment, storage medium, since it is corresponded to the methods disclosed in the examples, so be described relatively simple, correlation
Place is referring to method part illustration.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, the specific application and design constraint depending on technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond scope of the present application.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
A kind of image provided in an embodiment of the present invention removes misty rain method, apparatus, equipment and storage medium, this method comprises:
Data set is acquired first;The data set includes clear figure, and clearly schemes matched rain figure and mist figure;Then collected number is utilized
Deep learning CGAN network is trained according to collecting;Rain is carried out finally by trained deep learning CGAN network handles processing image
Mist processing.In view of the ambiguity that the generation of image is likely to result in weather misty rain, the application uses deep learning
CGAN network directly learns many rain figures and mist figure, can be effective when the image to be processed no matter inputted is rain figure or mist figure
Identification preferably utilizes the feature of whole figure, loses image feature information not, restored well by the correlation of image
Original image reaches good sharpening effect, reliability and accuracy is had more, and can handle most blurred picture, by image
Sharpening is easier, is easier.
Finally, it is to be noted that, herein, relational terms such as first and second and the like be used merely to by
One entity or operation are distinguished with another entity or operation, without necessarily requiring or implying these entities or operation
Between there are any actual relationship or orders.Moreover, the terms "include", "comprise" or its any other variant meaning
Covering non-exclusive inclusion, so that the process, method, article or equipment for including a series of elements not only includes that
A little elements, but also including other elements that are not explicitly listed, or further include for this process, method, article or
The intrinsic element of equipment.In the absence of more restrictions, the element limited by sentence "including a ...", is not arranged
Except there is also other identical elements in the process, method, article or apparatus that includes the element.
Misty rain method, apparatus, equipment and storage medium is gone to be described in detail image provided by the present invention above,
Used herein a specific example illustrates the principle and implementation of the invention, and the explanation of above embodiments is only used
In facilitating the understanding of the method and its core concept of the invention;At the same time, for those skilled in the art, according to the present invention
Thought, there will be changes in the specific implementation manner and application range, in conclusion the content of the present specification should not be construed as
Limitation of the present invention.
Claims (9)
1. a kind of image goes misty rain method characterized by comprising
Acquire data set;The data set includes clear figure, clearly schemes matched rain figure and mist figure with described;
Utilize the collected data set training deep learning CGAN network;
Misty rain is carried out by the trained deep learning CGAN network handles processing image to handle.
2. image according to claim 1 goes misty rain method, which is characterized in that utilize the collected data set training
Deep learning CGAN network, specifically includes:
Classify to the collected data set, is divided into training set and verifying collection;
The training set is input to deep learning CGAN network to be trained, is automatically adjusted in the deep learning CGAN network
The parameter of generator and arbiter;
The output of the deep learning CGAN network is observed in the training process as a result, judging whether to need by the verifying collection
Artificial modification hyper parameter and initialization value.
3. image according to claim 2 goes misty rain method, which is characterized in that the training set is input to deep learning
CGAN network is trained, and is automatically adjusted the parameter of generator and arbiter in the deep learning CGAN network, is specifically included:
Building condition generates deep learning CGAN network;The deep learning CGAN network include generator and arbiter the two
Model;
The training set is read, and the training set of reading and random noise Input matrix to the generator and are generated new
Image;
The new images that the generator generates are compared with the clear figure by the arbiter, according to comparing result meter
Calculate loss function;
Penalty values are calculated according to the loss function, automatic training and the ginseng for continuing to adjust the generator and the arbiter
Number, until the value relative minimum or epoch for making the loss function reach preset threshold.
4. image according to claim 3 goes misty rain method, which is characterized in that adjust the generator and the arbiter
Parameter, specifically include:
The parameter for adjusting the generator makes the new images generated in the generator close to the clear figure, not misjudged
Not;
The parameter for adjusting the arbiter keeps the differentiation probability of the clear figure in loss function bigger, raw in the generator
At new images differentiation probability it is bigger.
5. image according to claim 1 goes misty rain method, which is characterized in that acquisition data set specifically includes:
Collected clear figure is stored in the first file, will clearly scheme matched rain figure and mist figure the second file of deposit with described
Folder;
Identical name is carried out to the clear figure, the rain figure and the mist figure that are mutually matched;
The image size being stored into first file and second file is adjusted to unified picture format, as
Data set gathered in advance.
6. image according to claim 5 goes misty rain method, which is characterized in that with described clearly to scheme matched rain figure be pair
The clear figure carries out plus rain effect obtains;
With it is described clearly scheme matched mist figure be the clear figure is carried out plus fog effect obtain.
7. a kind of image removes misty rain device characterized by comprising
Dataset acquisition module, for acquiring data set;The data set includes clear figure, clearly schemes matched rain figure with described
With mist figure;
Network training module, for utilizing the collected data set training deep learning CGAN network;
Image processing module carries out at misty rain for handling image by the trained deep learning CGAN network handles
Reason.
8. a kind of image goes misty rain equipment, which is characterized in that including processor and memory, wherein the processor executes institute
It is realized when stating the computer program saved in memory as image as claimed in any one of claims 1 to 6 goes misty rain method.
9. a kind of computer readable storage medium, which is characterized in that for storing computer program, wherein the computer journey
It is realized when sequence is executed by processor as image as claimed in any one of claims 1 to 6 goes misty rain method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810688056.9A CN108986044A (en) | 2018-06-28 | 2018-06-28 | A kind of image removes misty rain method, apparatus, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810688056.9A CN108986044A (en) | 2018-06-28 | 2018-06-28 | A kind of image removes misty rain method, apparatus, equipment and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108986044A true CN108986044A (en) | 2018-12-11 |
Family
ID=64539364
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810688056.9A Pending CN108986044A (en) | 2018-06-28 | 2018-06-28 | A kind of image removes misty rain method, apparatus, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108986044A (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109859113A (en) * | 2018-12-25 | 2019-06-07 | 北京奇艺世纪科技有限公司 | Model generating method, image enchancing method, device and computer readable storage medium |
CN109886975A (en) * | 2019-02-19 | 2019-06-14 | 武汉大学 | It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network |
CN110009578A (en) * | 2019-03-13 | 2019-07-12 | 五邑大学 | A kind of image rain removing method and its system, device, storage medium |
CN110163813A (en) * | 2019-04-16 | 2019-08-23 | 中国科学院深圳先进技术研究院 | A kind of image rain removing method, device, readable storage medium storing program for executing and terminal device |
CN112507327A (en) * | 2020-08-31 | 2021-03-16 | 华南理工大学 | Weather detection and processing method based on machine learning |
CN113673361A (en) * | 2021-07-28 | 2021-11-19 | 东风汽车集团股份有限公司 | Rain and fog recognition method, sweeping system and computer-readable storage medium |
CN117196985A (en) * | 2023-09-12 | 2023-12-08 | 军事科学院军事医学研究院军事兽医研究所 | Visual rain and fog removing method based on deep reinforcement learning |
CN117455809A (en) * | 2023-10-24 | 2024-01-26 | 武汉大学 | Image mixed rain removing method and system based on depth guiding diffusion model |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106127702A (en) * | 2016-06-17 | 2016-11-16 | 兰州理工大学 | A kind of image mist elimination algorithm based on degree of depth study |
CN107798669A (en) * | 2017-12-08 | 2018-03-13 | 北京小米移动软件有限公司 | Image defogging method, device and computer-readable recording medium |
-
2018
- 2018-06-28 CN CN201810688056.9A patent/CN108986044A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106127702A (en) * | 2016-06-17 | 2016-11-16 | 兰州理工大学 | A kind of image mist elimination algorithm based on degree of depth study |
CN107798669A (en) * | 2017-12-08 | 2018-03-13 | 北京小米移动软件有限公司 | Image defogging method, device and computer-readable recording medium |
Non-Patent Citations (1)
Title |
---|
杨卫华 等: "《眼科人工智能》", 28 February 2018 * |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109859113A (en) * | 2018-12-25 | 2019-06-07 | 北京奇艺世纪科技有限公司 | Model generating method, image enchancing method, device and computer readable storage medium |
CN109886975A (en) * | 2019-02-19 | 2019-06-14 | 武汉大学 | It is a kind of that raindrop method and system is gone based on the image optimization processing for generating confrontation network |
CN110009578A (en) * | 2019-03-13 | 2019-07-12 | 五邑大学 | A kind of image rain removing method and its system, device, storage medium |
CN110163813A (en) * | 2019-04-16 | 2019-08-23 | 中国科学院深圳先进技术研究院 | A kind of image rain removing method, device, readable storage medium storing program for executing and terminal device |
CN110163813B (en) * | 2019-04-16 | 2022-02-01 | 中国科学院深圳先进技术研究院 | Image rain removing method and device, readable storage medium and terminal equipment |
CN112507327A (en) * | 2020-08-31 | 2021-03-16 | 华南理工大学 | Weather detection and processing method based on machine learning |
CN113673361A (en) * | 2021-07-28 | 2021-11-19 | 东风汽车集团股份有限公司 | Rain and fog recognition method, sweeping system and computer-readable storage medium |
CN117196985A (en) * | 2023-09-12 | 2023-12-08 | 军事科学院军事医学研究院军事兽医研究所 | Visual rain and fog removing method based on deep reinforcement learning |
CN117455809A (en) * | 2023-10-24 | 2024-01-26 | 武汉大学 | Image mixed rain removing method and system based on depth guiding diffusion model |
CN117455809B (en) * | 2023-10-24 | 2024-05-24 | 武汉大学 | Image mixed rain removing method and system based on depth guiding diffusion model |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108986044A (en) | A kind of image removes misty rain method, apparatus, equipment and storage medium | |
CN110544264B (en) | Temporal bone key anatomical structure small target segmentation method based on 3D deep supervision mechanism | |
CN106778867A (en) | Object detection method and device, neural network training method and device | |
CN108492271A (en) | A kind of automated graphics enhancing system and method for fusion multi-scale information | |
CN107004116B (en) | Method and apparatus for predicting face's attribute | |
CN110020714B (en) | Model training and data analysis method, device, equipment and storage medium | |
CN110245711A (en) | The SAR target identification method for generating network is rotated based on angle | |
CN108985133B (en) | Age prediction method and device for face image | |
CN105989341A (en) | Character recognition method and device | |
WO2019127102A1 (en) | Information processing method and apparatus, cloud processing device, and computer program product | |
CN109285154A (en) | A kind of method and device detecting the stone age | |
CN109919252A (en) | The method for generating classifier using a small number of mark images | |
CN108109212A (en) | A kind of historical relic restorative procedure, apparatus and system | |
JP2022500728A (en) | How to synthesize images based on conditional hostile generation networks and related equipment | |
CN116862931B (en) | Medical image segmentation method and device, storage medium and electronic equipment | |
CN107958472A (en) | PET imaging methods, device, equipment and storage medium based on sparse projection data | |
CN114299082A (en) | New coronary pneumonia CT image segmentation method, device and storage medium | |
CN104680190B (en) | Object detection method and device | |
JP6713422B2 (en) | Learning device, event detection device, learning method, event detection method, program | |
CN110874824B (en) | Image restoration method and device | |
CN104980725B (en) | Device and method for forming three-dimensional scence | |
CN116051813A (en) | Full-automatic intelligent lumbar vertebra positioning and identifying method and application | |
CN115641344A (en) | Method for segmenting optic disc image in fundus image | |
CN103208106B (en) | Detect the method for collimation side, device and x-ray imaging equipment | |
CN101000690A (en) | Method for improving image consistency, signal, computer readable medium and system |
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 |
Application publication date: 20181211 |
|
RJ01 | Rejection of invention patent application after publication |