CN109523474A - A kind of enhancement method of low-illumination image based on greasy weather degradation model - Google Patents
A kind of enhancement method of low-illumination image based on greasy weather degradation model Download PDFInfo
- Publication number
- CN109523474A CN109523474A CN201811220177.7A CN201811220177A CN109523474A CN 109523474 A CN109523474 A CN 109523474A CN 201811220177 A CN201811220177 A CN 201811220177A CN 109523474 A CN109523474 A CN 109523474A
- Authority
- CN
- China
- Prior art keywords
- image
- mist
- quasi
- low
- layer
- 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 28
- 230000015556 catabolic process Effects 0.000 title claims abstract description 15
- 238000006731 degradation reaction Methods 0.000 title claims abstract description 15
- 238000005286 illumination Methods 0.000 title claims abstract description 12
- 239000003595 mist Substances 0.000 claims abstract description 48
- 230000002708 enhancing effect Effects 0.000 claims abstract description 13
- 238000002834 transmittance Methods 0.000 claims description 16
- 230000000903 blocking effect Effects 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 239000000203 mixture Substances 0.000 claims description 4
- 238000003475 lamination Methods 0.000 claims description 2
- 238000013527 convolutional neural network Methods 0.000 abstract description 9
- 230000000007 visual effect Effects 0.000 abstract description 6
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 210000002569 neuron Anatomy 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 238000000889 atomisation Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 210000001525 retina Anatomy 0.000 description 1
- 229920006395 saturated elastomer Polymers 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
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
- G06T5/92—Dynamic range modification of images or parts thereof based on global image properties
-
- 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
- G06T2207/10024—Color image
-
- 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]
-
- 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
- G06T2207/20172—Image enhancement details
- G06T2207/20192—Edge enhancement; Edge preservation
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30192—Weather; Meteorology
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The present invention relates to a kind of enhancement method of low-illumination image based on greasy weather degradation model.This method inverts low-light (level) image first and obtains quasi- mist figure;Then transmissivity is solved using convolutional neural networks, and replaces global air light value using local atmosphere light value;Fog free images are finally restored based on atmospherical scattering model, reversion fog free images obtain the enhancing result of low-light (level) image.The method of the present invention had both effectively improved the brightness of low-light (level) image, in turn avoid apparent cross-color, it is over-exposed phenomena such as, image visual effect is good after enhancing.
Description
Technical field
The present invention relates to technical field of image processing, especially a kind of low-light (level) image enhancement based on greasy weather degradation model
Method.
Background technique
Under the low light conditions such as traditional night or backlight, acquired image visual effect is bad, and since contrast is low,
Effective information substantially reduces, and influences various subsequent analysis to some extent, such as target recognition and tracking etc. gives video monitoring
Many puzzlements are brought etc. numerous Video Applications.And image enhancement can adjust its Partial Feature according to the special scenes of given image,
Difference in enlarged image between different characteristic enhances interested feature, to play the visual effect for improving image, increases
The effect of effective information.Therefore, low-light (level) image enhancement problem is always the research hotspot of field of image enhancement.
Some traditional image enchancing methods have been demonstrated to be suitable for low-light (level) image enhancement problem.In Retinex method
For, this method is divided into two parts of incident image and reflected image using color constancy as theoretical basis, by piece image, leads to
Crossing reduces even influence of the removal incident image to reflected image, and decomposition obtains primary reflection component, it is non-uniform to eliminate reflection
It influences to achieve the effect that image enhancement.But this method colour cast is serious and noise-sensitive.For the color distortion phenomenon side MSRCR
Method (Multi-scale retinex with color restoration) MSR method (Multi-scale Retinex,
The enhancing of multiple dimensioned retina) on the basis of joined color and restore and color balancing, to each RGB component introduce color recieving because
Son adjusts the proportionate relationship of 3 Color Channels in original image, and the influence of color distortion is reduced while enhancing image.But draw
The color entered is restored parameter needs and is manually adjusted, and algorithm complexity is increased.2011, Dong etc. had found that low-light (level) image is asked
Anti- rear and Misty Image similitude.The discovery brings the thinking enhanced indirectly to low-light (level) image enhancement, i.e., by low-light (level)
The reversion figure of image is considered as quasi- mist image, after applying the defogging based on atomization degradation model on it, after reversion obtains enhancing
Image.Compared with true Misty Image, it is different at two aspects to intend mist image.Firstly, quasi- mist image usually has greatly
The bright areas of area.Since dark primary priori theoretical for large area bright areas and is not suitable for, therefore, it is impossible to using common
Mini-value filtering rough estimate intend mist image transmissivity.Secondly, quasi- mist image atmosphere brightness is usually relatively high.Work as atmosphere light
When value is close to 1, the quasi- mist image after defogging is partially dark, is easy to appear saturated phenomenon when inverting again.
In recent years, convolutional neural networks made breakthrough progress in classification and identification, its advantage lies in being able to face
Feature is adaptively extracted to particular problem, and the feature extracted has more identification.Learn quasi- mist using convolutional neural networks
Mapping relations between image and transmissivity are a kind of new to solve quasi- mist image large area bright areas transmissivity estimation problem
Feasible program.
Therefore, based on described above, the application learns the mapping between quasi- mist image and transmissivity based on convolutional neural networks
Relationship, in addition, introducing Steerable filter while global air light value is refined and optimizing to obtain atmosphere light figure.The method of the present invention exists
Have the advantages that be distorted less, noise sensitivity is low etc. while enhancing low-light (level) image, image visual effect is good after enhancing.
Summary of the invention
The purpose of the present invention is to provide a kind of enhancement method of low-illumination image based on greasy weather degradation model, this method exists
Have the advantages that be distorted less, noise sensitivity is low etc. while enhancing low-light (level) image, image visual effect is good after enhancing.
To achieve the above object, the technical scheme is that a kind of low-light (level) image based on greasy weather degradation model increases
Strong method, comprising the following steps:
Step S1, the low-light (level) image R of input is negated to obtain quasi- mist figure I;
Step S2, the grayscale image I of quasi- mist figure I is soughtgray;
Step S3, transmissivity prediction model of the training based on CNN acquires the thick transmittance figure T of quasi- mist figure I;
Step S4, the grayscale image I obtained with step S2grayT is modified as navigational figure, the transmission refined
Rate figure T ';
Step S5, the quasi- mist figure I piecemeal obtained to step S1, every piece of maximum value taken in three Color Channels is as part
Air light value obtains thick atmosphere light figure a;
Step S6, the grayscale image I obtained with step S2grayA is modified as navigational figure, the atmosphere refined
Light figure a '.
Step S7, ask quasi- mist figure I corresponding with the atmosphere light figure a ' that step S4 obtained transmittance figure T ' and step S6 are obtained
Fog free images I ';
Step S8, fog free images I ' is negated to obtain final low-light (level) enhancing image R '.
In an embodiment of the present invention, in the step S3, the transmissivity prediction model based on CNN of training, specifically
Setting are as follows: the input of network is the image of 3 × 16 × 16 sizes, and first layer convolutional layer is made of 16 characteristic patterns, and filter is big
Small is 5 × 5;Slice layers of the second layer are made of 4 characteristic patterns;Eltwise layers of third layer are made of 4 characteristic patterns;4th layer of volume
Lamination is made of 48 characteristic patterns, and filter size is respectively 3 × 3,5 × 5,7 × 7;Layer 5 pond layer is by 48 characteristic patterns
Composition, filter size are 7 × 7;Layer 6 convolutional layer is made of a characteristic pattern, and filter size is 6 × 6;The last layer
It is BreLU layers, exports as final transmissivity predicted value.
In an embodiment of the present invention, in the step S4, by the grayscale image I of quasi- mist figure IgrayAs navigational figure;
The transmittance figure T that quasi- mist figure I is generated by the step S3 is as input picture;Utilize the mean value of navigational figure itself, side
Mean value, variance operation between difference operation and navigational figure and input picture retain the content of restored image, and described in transmitting
The smooth edge details of quasi- mist figure give output image, to eliminate the blocking artifact of image, obtain the transmittance figure of transmissivity optimization
T′。
In an embodiment of the present invention, in the step S5, specifically comprise the following steps:
Step S51: quasi- mist figure I is divided into 16 × 16 fritter;
Step S52: each 16 × 16 fritter is sought by the maximum values of three Color Channels is obtained as local atmosphere light value
To thick atmosphere light figure a.
In an embodiment of the present invention, in the step S6, by the grayscale image I of quasi- mist figure IgrayAs navigational figure;It will
The transmittance figure a that the step S5 is generated is as input picture;Utilize the mean value of navigational figure itself, variance operation and guidance
Mean value, variance operation between image and input picture retain the content of restored image, and transmit the smooth side of the quasi- mist figure
Edge details gives output image, to eliminate the blocking artifact of image, obtains the atmosphere light figure a ' of transmissivity optimization.
In an embodiment of the present invention, in the step S7, the specific formula of the corresponding fog free images I ' of quasi- mist figure I is sought
Are as follows:
Wherein, x is pixel.
Compared to the prior art, the invention has the following advantages: the present invention provides one kind to be based on greasy weather degeneration mould
The enhancement method of low-illumination image of type, this method had both effectively improved the brightness of low-light (level) image, in turn avoided apparent color
Phenomena such as being distorted, is over-exposed, image visual effect is good after enhancing.
Detailed description of the invention
Fig. 1 is the flow chart of the image enchancing method based on greasy weather degradation model in the present invention.
Fig. 2 is the network structure of the transmissivity prediction model based on CNN in the present invention.
Specific embodiment
With reference to the accompanying drawing, technical solution of the present invention is specifically described.
The present invention provides a kind of enhancement method of low-illumination image based on greasy weather degradation model, as shown in Figure 1, specific packet
Include following steps:
Step S1, the low-light (level) image R of input is negated to obtain quasi- mist figure I;
Step S2, the grayscale image I of quasi- mist figure I is soughtgray;
Step S3, transmissivity prediction model of the training based on CNN acquires the thick transmittance figure T of quasi- mist figure I;
Step S4, the grayscale image I obtained with step S2grayT is modified as navigational figure, the transmission refined
Rate figure T ';
Step S5, the quasi- mist figure I piecemeal obtained to step S1, every piece of maximum value taken in three Color Channels is as part
Air light value obtains thick atmosphere light figure a;
Step S6, the grayscale image I obtained with step S2grayA is modified as navigational figure, the atmosphere refined
Light figure a '.
Step S7, ask quasi- mist figure I corresponding with the atmosphere light figure a ' that step S4 obtained transmittance figure T ' and step S6 are obtained
Fog free images I ';
Step S8, fog free images I ' is negated to obtain final low-light (level) enhancing image R '.
In the present embodiment, in the step S1, the specific formula for calculation of quasi- mist figure I is sought are as follows:
I (x)=1-R (x)
Wherein, x is pixel.
In the present embodiment, in the step S3, the transmissivity prediction model based on CNN of training is as shown in Fig. 2, tool
Body setting are as follows: the input of network is the image of 3 × 16 × 16 sizes, the convolutional layer that first layer is made of 16 characteristic patterns, often
A neuron is connected with one 5 × 5 of input picture neighborhood, therefore the size of each characteristic pattern is 12 × 12;The second layer
The result of first layer is cut into 4 parts by the Slice layer being made of 44 × 12 × 12 characteristic patterns;Third layer is by 4
The Eltwise layer of a characteristic pattern composition, merges into 14 × 12 by being maximized method for 44 × 12 × 12 characteristic patterns
× 12 characteristic pattern with pixel maximum;The 4th layer of convolutional layer being made of 48 characteristic patterns, it is 3 that size, which is respectively adopted,
× 3,5 × 5,7 × 7 filter carries out convolution operation to input respectively, carries out multi-scale expression;Layer 5 is by 48 6 × 6
The pond layer of characteristic pattern composition uses the filter progress pondization that size is 7 × 7 to operate for learning local extremum feature, from
Value of the maximum value as current whole region is chosen in 7 × 7 region;The convolution that layer 6 is made of a characteristic pattern
Layer, for characteristic pattern to be converted to 1 × 1 characteristic value information, each neuron and one 6 × 6 of input picture neighborhood phase
Connection, therefore the size of each characteristic pattern is 1 × 1;The last layer is BreLU layers, is carried out to the characteristic pattern of 1 × 1 size non-thread
Property variation, gained functional value is constrained in [0,1] range, is exported as final transmissivity predicted value.
In the present embodiment, in the step S4, by the grayscale image I of quasi- mist figure IgrayAs navigational figure;By quasi- mist
Transmittance figure T of the I by step S3 generation is schemed as input picture;Utilize the mean value of navigational figure itself, variance operation
And the mean value between navigational figure and input picture, variance operation retain the content of restored image, and transmit the quasi- mist figure
Smooth edge details give output image, to eliminate the blocking artifact of image, obtain the transmittance figure T ' of transmissivity optimization.
In the present embodiment, in the step S5, specifically comprise the following steps:
Step S51: quasi- mist figure I is divided into 16 × 16 fritter;
Step S52: each 16 × 16 fritter is sought by the maximum values of three Color Channels is obtained as local atmosphere light value
To thick atmosphere light figure a, specific formula for calculation are as follows:
Wherein, x is pixel, and i indicates the label of fritter.
In the present embodiment, in the step S6, by the grayscale image I of quasi- mist figure IgrayAs navigational figure;By the step
The transmittance figure a that rapid S5 is generated is as input picture;Using the mean value of navigational figure itself, variance operation and navigational figure with
Mean value, variance operation between input picture retain the content of restored image, and transmit the smooth edge details of the quasi- mist figure
Output image is given, to eliminate the blocking artifact of image, obtains the atmosphere light figure a ' of transmissivity optimization.
In the present embodiment, in the step S7, the specific formula of the corresponding fog free images I ' of quasi- mist figure I is sought are as follows:
Wherein, x is pixel.
In the present embodiment, in step s 8: final low-light (level) being asked to enhance image R ', specific formula are as follows:
R (x)=1-I (x)
Wherein, x is pixel.
The above are preferred embodiments of the present invention, all any changes made according to the technical solution of the present invention, and generated function is made
When with range without departing from technical solution of the present invention, all belong to the scope of protection of the present invention.
Claims (6)
1. a kind of enhancement method of low-illumination image based on greasy weather degradation model, which comprises the following steps:
Step S1, the low-light (level) image R of input is negated to obtain quasi- mist figure I;
Step S2, the grayscale image I of quasi- mist figure I is soughtgray;
Step S3, transmissivity prediction model of the training based on CNN acquires the thick transmittance figure T of quasi- mist figure I;
Step S4, the grayscale image I obtained with step S2grayT is modified as navigational figure, the transmittance figure refined
T′;
Step S5, the quasi- mist figure I piecemeal obtained to step S1, every piece of maximum value taken in three Color Channels is as local atmosphere
Light value obtains thick atmosphere light figure a;
Step S6, the grayscale image I obtained with step S2grayA is modified as navigational figure, the atmosphere light figure refined
a′。
Step S7, the corresponding nothing of quasi- mist figure I is sought with the atmosphere light figure a ' that step S4 obtained transmittance figure T ' and step S6 are obtained
Mist image I ';
Step S8, fog free images I ' is negated to obtain final low-light (level) enhancing image R '.
2. a kind of enhancement method of low-illumination image based on greasy weather degradation model according to claim 1, which is characterized in that
In the step S3, the transmissivity prediction model based on CNN of training, be specifically configured to: the input of network is 3 × 16 × 16
The image of size, first layer convolutional layer are made of 16 characteristic patterns, and filter size is 5 × 5;Slice layers of the second layer by 4 spies
Sign figure composition;Eltwise layers of third layer are made of 4 characteristic patterns;4th layer of convolutional layer is made of 48 characteristic patterns, and filter is big
Small is respectively 3 × 3,5 × 5,7 × 7;Layer 5 pond layer is made of 48 characteristic patterns, and filter size is 7 × 7;Layer 6 volume
Lamination is made of a characteristic pattern, and filter size is 6 × 6;The last layer is BreLU layers, is exported as the prediction of final transmissivity
Value.
3. a kind of enhancement method of low-illumination image based on greasy weather degradation model according to claim 1, which is characterized in that
In the step S4, by the grayscale image I of quasi- mist figure IgrayAs navigational figure;Quasi- mist figure I is generated by the step S3
Transmittance figure T as input picture;Utilize the mean value of navigational figure itself, variance operation and navigational figure and input picture
Between mean value, variance operation retain the content of restored image, and transmit the smooth edge details of the quasi- mist figure and give output figure
Picture obtains the transmittance figure T ' of transmissivity optimization to eliminate the blocking artifact of image.
4. a kind of enhancement method of low-illumination image based on greasy weather degradation model according to claim 1, which is characterized in that
In the step S5, specifically comprise the following steps:
Step S51: quasi- mist figure I is divided into 16 × 16 fritter;
Step S52: each 16 × 16 fritter is sought by the maximum values of three Color Channels is obtained slightly as local atmosphere light value
Atmosphere light figure a.
5. a kind of enhancement method of low-illumination image based on greasy weather degradation model according to claim 1, which is characterized in that
In the step S6, by the grayscale image I of quasi- mist figure IgrayAs navigational figure;The transmittance figure a that the step S5 is generated makees
For input picture;Utilize the mean value between the mean value of navigational figure itself, variance operation and navigational figure and input picture, side
Difference operation retains the content of restored image, and transmits the smooth edge details of the quasi- mist figure and give output image, to eliminate figure
The blocking artifact of picture obtains the atmosphere light figure a ' of transmissivity optimization.
6. a kind of enhancement method of low-illumination image based on greasy weather degradation model according to claim 1, which is characterized in that
In the step S7, the specific formula of the corresponding fog free images I ' of quasi- mist figure I is sought are as follows:
Wherein, x is pixel.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811220177.7A CN109523474A (en) | 2018-10-19 | 2018-10-19 | A kind of enhancement method of low-illumination image based on greasy weather degradation model |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201811220177.7A CN109523474A (en) | 2018-10-19 | 2018-10-19 | A kind of enhancement method of low-illumination image based on greasy weather degradation model |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109523474A true CN109523474A (en) | 2019-03-26 |
Family
ID=65770582
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201811220177.7A Pending CN109523474A (en) | 2018-10-19 | 2018-10-19 | A kind of enhancement method of low-illumination image based on greasy weather degradation model |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109523474A (en) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109978962A (en) * | 2019-04-09 | 2019-07-05 | 广州市交通高级技工学校(广州市交通技师学院) | A kind of low contrast indicating value digital image recognition method towards the calibrating of darkroom illumination photometer |
CN110136079A (en) * | 2019-05-05 | 2019-08-16 | 长安大学 | Image defogging method based on scene depth segmentation |
CN112381897A (en) * | 2020-11-16 | 2021-02-19 | 西安电子科技大学 | Low-illumination image enhancement method based on self-coding network structure |
CN112529802A (en) * | 2020-12-09 | 2021-03-19 | 电子科技大学 | Atmospheric scattering degraded image recovery method based on scattering coefficient ratio estimation |
CN112614063A (en) * | 2020-12-18 | 2021-04-06 | 武汉科技大学 | Image enhancement and noise self-adaptive removal method for low-illumination environment in building |
CN112819707A (en) * | 2021-01-15 | 2021-05-18 | 电子科技大学 | End-to-end anti-blocking effect low-illumination image enhancement method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8611691B2 (en) * | 2009-07-31 | 2013-12-17 | The United States Of America As Represented By The Secretary Of The Army | Automated video data fusion method |
CN106600560A (en) * | 2016-12-22 | 2017-04-26 | 福州大学 | Image defogging method for automobile data recorder |
CN108133461A (en) * | 2017-12-07 | 2018-06-08 | 中国航空工业集团公司西安航空计算技术研究所 | A kind of real-time video low-light (level) Enhancement Method |
-
2018
- 2018-10-19 CN CN201811220177.7A patent/CN109523474A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8611691B2 (en) * | 2009-07-31 | 2013-12-17 | The United States Of America As Represented By The Secretary Of The Army | Automated video data fusion method |
CN106600560A (en) * | 2016-12-22 | 2017-04-26 | 福州大学 | Image defogging method for automobile data recorder |
CN108133461A (en) * | 2017-12-07 | 2018-06-08 | 中国航空工业集团公司西安航空计算技术研究所 | A kind of real-time video low-light (level) Enhancement Method |
Non-Patent Citations (1)
Title |
---|
余春艳 等: "应用雾天退化模型的低照度图像增强", 《中国图像图形学报》 * |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109978962A (en) * | 2019-04-09 | 2019-07-05 | 广州市交通高级技工学校(广州市交通技师学院) | A kind of low contrast indicating value digital image recognition method towards the calibrating of darkroom illumination photometer |
CN109978962B (en) * | 2019-04-09 | 2022-05-17 | 广州市交通高级技工学校(广州市交通技师学院) | Low-contrast indicating value image intelligent identification method for darkroom illuminometer calibration |
CN110136079A (en) * | 2019-05-05 | 2019-08-16 | 长安大学 | Image defogging method based on scene depth segmentation |
CN112381897A (en) * | 2020-11-16 | 2021-02-19 | 西安电子科技大学 | Low-illumination image enhancement method based on self-coding network structure |
CN112381897B (en) * | 2020-11-16 | 2023-04-07 | 西安电子科技大学 | Low-illumination image enhancement method based on self-coding network structure |
CN112529802A (en) * | 2020-12-09 | 2021-03-19 | 电子科技大学 | Atmospheric scattering degraded image recovery method based on scattering coefficient ratio estimation |
CN112529802B (en) * | 2020-12-09 | 2021-07-20 | 电子科技大学 | Atmospheric scattering degraded image recovery method based on scattering coefficient ratio estimation |
CN112614063A (en) * | 2020-12-18 | 2021-04-06 | 武汉科技大学 | Image enhancement and noise self-adaptive removal method for low-illumination environment in building |
CN112614063B (en) * | 2020-12-18 | 2022-07-01 | 武汉科技大学 | Image enhancement and noise self-adaptive removal method for low-illumination environment in building |
CN112819707A (en) * | 2021-01-15 | 2021-05-18 | 电子科技大学 | End-to-end anti-blocking effect low-illumination image enhancement method |
CN112819707B (en) * | 2021-01-15 | 2022-05-03 | 电子科技大学 | End-to-end anti-blocking effect low-illumination image enhancement method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110148095B (en) | Underwater image enhancement method and enhancement device | |
CN109523474A (en) | A kind of enhancement method of low-illumination image based on greasy weather degradation model | |
Xu et al. | Fast image dehazing using improved dark channel prior | |
CN108765342A (en) | A kind of underwater image restoration method based on improvement dark | |
Gao et al. | Sand-dust image restoration based on reversing the blue channel prior | |
CN109087254B (en) | Unmanned aerial vehicle aerial image haze sky and white area self-adaptive processing method | |
CN109816608B (en) | Low-illumination image self-adaptive brightness enhancement method based on noise suppression | |
CN107977942B (en) | Single image restoration method based on multi-focus fusion | |
CN112116536A (en) | Low-illumination image enhancement method and system | |
CN111476732B (en) | Image fusion and denoising method and system | |
CN107067375A (en) | A kind of image defogging method based on dark channel prior and marginal information | |
CN106023110B (en) | A kind of Hi-Fi image defogging method | |
CN104899836A (en) | Foggy image enhancing device and method based on near infrared multispectral imaging | |
CN103839245B (en) | The Retinex colour-image reinforcing method at night of Corpus--based Method rule | |
CN108648160B (en) | Underwater sea cucumber image defogging enhancement method and system | |
CN104392425A (en) | Face based automatic contrast adjusting image enhancing method | |
Dar et al. | An enhanced adaptive histogram equalization based local contrast preserving technique for HDR images | |
CN103226809A (en) | Image haze removal device and image haze removal method | |
CN107025641A (en) | Image interfusion method based on Analysis of Contrast | |
CN106709876A (en) | Optical remote sensing image defogging method based on the principle of dark pixel | |
CN112308793A (en) | Novel method for enhancing contrast and detail of non-uniform illumination image | |
Ma et al. | Image adaptive contrast enhancement for low-illumination lane lines based on improved Retinex and guided filter | |
CN110738624A (en) | area self-adaptive image defogging system and method | |
CN110992287A (en) | Method for clarifying non-uniform illumination video | |
CN113379631B (en) | Image defogging method and device |
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: 20190326 |