CN114897710A - Video enhancement method under low-illumination environment - Google Patents
Video enhancement method under low-illumination environment Download PDFInfo
- Publication number
- CN114897710A CN114897710A CN202210353200.XA CN202210353200A CN114897710A CN 114897710 A CN114897710 A CN 114897710A CN 202210353200 A CN202210353200 A CN 202210353200A CN 114897710 A CN114897710 A CN 114897710A
- Authority
- CN
- China
- Prior art keywords
- image
- mapping
- night
- algorithm
- value
- 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
- 238000005286 illumination Methods 0.000 title claims abstract description 23
- 238000000034 method Methods 0.000 title claims abstract description 15
- 238000003384 imaging method Methods 0.000 claims abstract description 7
- 230000000694 effects Effects 0.000 claims abstract description 6
- 238000013507 mapping Methods 0.000 claims description 33
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000003044 adaptive effect Effects 0.000 claims description 4
- 230000001186 cumulative effect Effects 0.000 claims description 4
- 238000005315 distribution function Methods 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 3
- 230000000903 blocking effect Effects 0.000 claims description 2
- 239000003086 colorant Substances 0.000 claims description 2
- 230000000007 visual effect Effects 0.000 claims description 2
- 230000002411 adverse Effects 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- 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
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/40—Image enhancement or restoration using histogram techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- 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
- G06T5/94—Dynamic range modification of images or parts thereof based on local image properties, e.g. for local contrast enhancement
-
- 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
- G06T2207/10016—Video; Image sequence
-
- 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
- G06T2207/20004—Adaptive image processing
-
- 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
- G06T2207/20021—Dividing image into blocks, subimages or windows
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention belongs to the technical field of video enhancement, and particularly relates to a video enhancement method in a low-illumination environment. Aiming at the adverse driving environment of rain, fog and low illumination at night, the functions of defogging and low illumination are enhanced in real time, and a clear view is obtained. In the aspect of defogging at night, the experimental result shows that the algorithm has good defogging effect, and people and more information in the dark environment in the image can be seen. The algorithm improves the image contrast according to the imaging characteristics of the night image, and the image processing result is relatively clear.
Description
Technical Field
The invention aims at the adverse driving environment of rain, fog and low illumination at night, and obtains a clear view by the real-time defogging and low illumination enhancement functions.
Background
In low-illumination image enhancement, common function mapping is easy to cause insufficient or excessive enhancement of an image while enhancing a low-illumination image. In the aspect of the night defogging mode, the SSR algorithm does not improve the contrast of the image because the SSR algorithm is in a low-illumination environment at night; the night imaging model of the Hommine algorithm is different from the daytime condition, and the defogging effect of the Hommine algorithm is not ideal due to the existence of multiple night light sources and the extremely complex night physical imaging model; the method of Li et al cannot be satisfactory for enhancing white pictures such as water surface. Meanwhile, the surrounding phenomenon of a dark area is obvious near the lamplight in the train diagram, the method overcomes the defects, and clear video images can be obtained in the adverse driving environment of rain, fog and low illumination at night.
Disclosure of Invention
The invention aims to solve the technical problem of a prediction method of functions of defogging and low-illumination enhancement in real time aiming at adverse driving environments of rain, fog and low-illumination at night.
Drawings
FIG. 1: and (5) after the circular arc function mapping algorithm, obtaining a gray level change graph. The horizontal axis represents the gray scale range of the original image and the vertical axis represents the warp function f 1 (x) The mapped gray scale range.
FIG. 2: and 4. a night defogging algorithm flow chart.
FIG. 3: the schematic diagram of the Retinex algorithm defogging is shown, wherein R (x, y) represents the light reflected by an actual object, namely the real color of the object, namely an image needing to be restored; l (x, y) represents a light ray irradiated from the light source to the surface of the object.
Detailed Description
1. The circular arc curve mapping algorithm:
2. considering that the arc curve mapping has the characteristics of equal range mapping and good arc smoothness, the invention provides an arc function mapping algorithm, and the expression is as follows:
l is the maximum value of the gray scale, generally 255, is the range of pixel values, and generally ranges from 0 to 255; f. of 1 (x) Is the gray value after mapping. Derivative f '(x) for any mapping function f (x), when f' (x)>1, the pixel value range is expanded; and 0<f'(x)<When 1, the pixel value range is reduced, the image hierarchical structure is easy to be unclear, and the image detail content is covered. Take L-255 as an example, when f' 4 When 1, x is 74.68. That is, x ≦ 74, the grayscale value is stretched and the image is enhanced; x is the number of>At 74, the gray scale range is compressed, and the gray scale change diagram is as shown in fig. 1.
5. Adaptive function mapping algorithm: combining the smoothness of the slope of the arc curve, a new determination method is provided, and the calculation formula is as follows:
7. wherein, for adjusting the parameters, the cumulative distribution function of the image histogram is represented, and the calculation formula is as follows:
9. from the above analysis, the calculation steps of the algorithm are shown in table 4.2:
10. the low-illuminance image I is input.
11. The cumulative distribution function of the RGB color channels is calculated using equation (3).
12. The gamma values for the RGB color channels are calculated using equation (2).
13. And (3) substituting the gamma value into the formula (3) to obtain the self-adaptive gray mapping curves of the three RGB color channels.
14. And updating the pixel values of the three color channels according to the mapping relation of the gray scale, and finally combining the three color channels to obtain an image J.
15. The enhanced image J is output.
16. Defogging at night: the night defogging process is shown in figure 2.
17. And Gaussian low-pass filtering treatment: homomorphic filtering takes an illumination-reflection model as a theoretical basis for image processing [90] According to the theory, a mathematical formula is adopted for transformation and is transformed into a frequency domain space, so that high-frequency components formed by strong illumination in the space are conveniently eliminated, the problem of prominent illumination is solved, and the aim of balancing image light is fulfilled.
Retinex algorithm defogging: a night fogging image P (x, y) can be expressed by the formula (4-4), and is schematically shown in FIG. 3. Namely:
19.P(x,y)=L(x,y)·R(x,y) (4-4)
20. in the above formula, R (x, y) represents the light reflected by the actual object, i.e. the true color of the object, i.e. the image to be restored; l (x, y) represents the light rays from the light source that strike the surface of the object and should be subtracted as much as possible during the calculation.
The CLAHE algorithm: the method comprises the following steps: inputting a night fogging image as an original image, dividing the night fogging image into M × N sub-blocks, and then dividing the sub-blocks into M × N sub-blocksCounting the number n corresponding to each pixel value i in each sub block i (ii) a Step two: each subblock is clipped with a custom threshold β, i.e. N i The values exceeding β are equally distributed to each gray level. Step three: the number of the pixel values i after the redistribution is N i . And mapping the gray value i of each sub-block to T (i) to obtain a new image. The mapping formula is as follows:wherein l max Maximum for each sub-block pixel. Step four: the center point of each sub-block is taken as a sample point, and then each pixel is interpolated by using the formula (5-15) so as to eliminate the problem of unsmooth between each sub-block caused by blocking. And the pixel value after interpolation calculation is the pixel value of the output image.
22.
23. Where f (x, y) represents the value of each pixel, and f (Q) 11 )、f(Q 21 )、f(Q 12 ) And f (Q) 22 ) Respectively representing the values of the pixels at the center points of four adjacent sub-blocks.
24. Color correction: the white balance theory holds that even in the colorful colors, the pixel mean values of the RGB individual color layers of the image are almost equal, i.e.: r is p =G p =B p Wherein R is p 、G p And B p The representation is the average of three color layers. And then calculating the ratio of each color layer according to the formula (6), wherein the formula is as follows:
26. wherein, the K value calculation formula is as follows:then according to VonAnd a Kries diagonal model, and a new value is obtained again.
27. In terms of low-illumination image enhancement: the circular arc curve mapping algorithm can well improve the gray scale range of the image, so that the image is clearer and has more distinct levels, and a good overall visual effect is obtained; the overall effect of the adaptive function mapping algorithm in the aspects of objective index average gradient, information entropy, contrast ratio and the like is superior to that of the algorithms in the text.
28. In terms of nighttime defogging: the experimental result shows that the algorithm has good defogging effect, and people and more information in the dark environment in the figure can be seen. The algorithm of the section improves the image contrast according to the imaging characteristics of the night image, and the image processing result is clearer.
Claims (2)
1. A video enhancement technique for rain, fog, and low-light environments at night, characterized by: the method comprises the following steps:
the gray level mapping is mainly improved, the image enhancement essence based on the gray level mapping is that the color range which can be distinguished by human eyes is smaller, but after the image is converted into a digital image, the gray level range which can be identified by digital equipment is wider, the gray level mapping essence is mapping, the function transformation is used for further enhancing the layering of the gray level, more details are highlighted, and the image content is enriched.
2. The video enhancement technique for rain, fog and night time low light environments of claim 1, wherein:
1) the SSR algorithm obtains an output image by removing an incident component, and in the aspect of low-illumination image enhancement, common function mapping easily causes insufficient or excessive enhancement of the image while enhancing the low-illumination image; in the aspect of the night defogging mode, the SSR algorithm does not improve the contrast of the image because the SSR algorithm is in a low-illumination environment at night; the night imaging model of the Homing algorithm is different from the daytime condition, and the night physical imaging model is extremely complex due to the existence of multiple night light sources;
2) arc curve mapping algorithm:
3) considering that the arc curve mapping has the characteristics of equal range mapping and good arc smoothness, the invention provides an arc function mapping algorithm, and the expression is as follows:
4) l is the maximum value of the gray scale, generally 255, and is a pixel value range, generally 0-255; f. of 1 (x) The gray value after mapping; for the derivative f '(x) of the arbitrary mapping function f (x), the pixel value range is expanded when f' (x) > 1; when f (x) is less than 1 and 0 is less than f, the range of pixel values is reduced, the image hierarchical structure is not clear easily, and the detail content of the image is covered; take L-255 as an example, when f' 4 When 1, x is 74.68; that is, x ≦ 74, the grayscale value is stretched and the image is enhanced; when x is larger than 74, the gray scale range is compressed;
5) adaptive function mapping algorithm: combining the smoothness of the slope of the arc curve, a new determination method is provided, and the calculation formula is as follows:
7) wherein, for adjusting the parameters, the cumulative distribution function of the image histogram is represented, and the calculation formula is as follows:
9) from the above analysis, the computational steps of the algorithm are shown in table 4.2:
10) inputting a low-illumination image I;
11) calculating a cumulative distribution function of the RGB color channels using formula (3);
12) calculating gamma values of the RGB color channels using formula (2);
13) substituting the gamma value into an expression (3) to obtain self-adaptive gray mapping curves of three RGB color channels;
14) updating pixel values of the three color channels according to the mapping relation of the gray levels, and finally combining the three color channels to obtain an image J;
15) outputting an enhanced image J;
16) defogging at night: the night defogging flow chart is shown in figure 2;
17) and Gaussian low-pass filtering treatment: homomorphic filtering takes an irradiation-reflection model as the theoretical basis of image processing, adopts mathematical formula transformation according to the theory, transforms the model into a frequency domain space, conveniently eliminates high-frequency components formed by strong illumination in the space, further solves the problem of prominent illumination and achieves the aim of balancing image light;
18) a night fogging image P (x, y) can be expressed by a formula (4); namely:
19)P(x,y)=L(x,y)·R(x,y) (4)
20) in the above formula, R (x, y) represents the light reflected by the actual object, i.e. the true color of the object, i.e. the image to be restored; l (x, y) represents the light rays irradiated to the surface of the object by the light source, and should be subtracted as much as possible in the calculation process;
21) the CLAHE algorithm: the method comprises the following steps: inputting a night foggy image as an original image, dividing the night foggy image into M multiplied by N subblocks, and counting the number N corresponding to each pixel value i in each subblock i (ii) a Step two: each subblock is clipped with a custom threshold β, i.e. N i The values exceeding β are equally distributed to each gray level; step three: the number of the pixel values i after the redistribution is N i (ii) a Mapping the gray value i of each sub-block to T (i) to obtain a new image, wherein the mapping formula is as follows:wherein l max Maximum value of pixel for each sub-block; step four: taking the central point of each sub-block as a sample point, and then performing interpolation calculation on each pixel by using a formula (5) so as to eliminate the unsmooth problem among the sub-blocks caused by blocking, wherein the pixel value after interpolation calculation is the pixel value of the output image;
22) color correction: the white balance theory holds that even in the colorful colors, the pixel mean values of the RGB individual color layers of the image are almost equal, i.e.: r p =G p =B p Wherein R is p 、G p And B p The expression is the average value of each color layer, and the ratio of each color layer is calculated according to the formula (6), which is as follows:
23)
where f (x, y) represents the value of each pixel, and f (Q) 11 )、f(Q 21 )、f(Q 12 ) And f (Q) 22 ) Respectively showing the values of the pixels at the center points of four adjacent sub-blocks,
24) wherein, the K value calculation formula is as follows:then, according to a Von Kries diagonal model, a new value is obtained again;
25) in terms of low-illumination image enhancement: the circular arc curve mapping algorithm can well improve the gray scale range of the image, so that the image is clearer and has more distinct levels, and a good overall visual effect is obtained; the overall effect of the adaptive function mapping algorithm in the aspects of objective index average gradient, information entropy, contrast ratio and the like is superior to that of the algorithms in the text;
26) in terms of nighttime defogging: the experimental result shows that the algorithm has good defogging effect, and people and more information in the dark environment in the figure can be seen; the algorithm of the section improves the image contrast according to the imaging characteristics of the night image, and the image processing result is relatively clear.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210353200.XA CN114897710A (en) | 2022-04-06 | 2022-04-06 | Video enhancement method under low-illumination environment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210353200.XA CN114897710A (en) | 2022-04-06 | 2022-04-06 | Video enhancement method under low-illumination environment |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114897710A true CN114897710A (en) | 2022-08-12 |
Family
ID=82716328
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210353200.XA Pending CN114897710A (en) | 2022-04-06 | 2022-04-06 | Video enhancement method under low-illumination environment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114897710A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115587950A (en) * | 2022-11-03 | 2023-01-10 | 昆山腾云达信息咨询技术服务中心(有限合伙) | Low-light-level enhanced color recovery method |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106251300A (en) * | 2016-07-26 | 2016-12-21 | 华侨大学 | A kind of quick night of based on Retinex Misty Image restored method |
GB202006954D0 (en) * | 2019-05-14 | 2020-06-24 | Univ Beijing Science & Technology | Underwater image enhancement method and enhancement device |
CN113947535A (en) * | 2020-07-17 | 2022-01-18 | 四川大学 | Low-illumination image enhancement method based on illumination component optimization |
-
2022
- 2022-04-06 CN CN202210353200.XA patent/CN114897710A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106251300A (en) * | 2016-07-26 | 2016-12-21 | 华侨大学 | A kind of quick night of based on Retinex Misty Image restored method |
GB202006954D0 (en) * | 2019-05-14 | 2020-06-24 | Univ Beijing Science & Technology | Underwater image enhancement method and enhancement device |
CN113947535A (en) * | 2020-07-17 | 2022-01-18 | 四川大学 | Low-illumination image enhancement method based on illumination component optimization |
Non-Patent Citations (1)
Title |
---|
程鸿: "车辆安全预警系统与夜间视频增强技术研究" * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115587950A (en) * | 2022-11-03 | 2023-01-10 | 昆山腾云达信息咨询技术服务中心(有限合伙) | Low-light-level enhanced color recovery method |
CN115587950B (en) * | 2022-11-03 | 2023-09-26 | 昆山腾云达信息咨询技术服务中心(有限合伙) | Low-light-level enhanced color recovery method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110148095B (en) | Underwater image enhancement method and enhancement device | |
CN106157267B (en) | Image defogging transmissivity optimization method based on dark channel prior | |
CN104021532B (en) | A kind of image detail enhancement method of infrared image | |
CN102789635B (en) | Image enhancement method and image enhancement device | |
CN109447917B (en) | Remote sensing image haze eliminating method based on content, characteristics and multi-scale model | |
CN107909562A (en) | A kind of Fast Image Fusion based on Pixel-level | |
CN110706172B (en) | Low-illumination color image enhancement method based on adaptive chaotic particle swarm optimization | |
CN109816608B (en) | Low-illumination image self-adaptive brightness enhancement method based on noise suppression | |
CN107862672B (en) | Image defogging method and device | |
Liu et al. | Low-light video image enhancement based on multiscale retinex-like algorithm | |
CN109003238B (en) | Image haze removal method based on model, histogram and gray level enhancement | |
CN102208101A (en) | Self-adaptive linearity transformation enhancing method of infrared image | |
CN106971379A (en) | A kind of underwater picture Enhancement Method merged based on stratified calculation | |
CN115660964A (en) | Nighttime road image enhancement method based on threshold partition weighted brightness component | |
CN113313639A (en) | Image enhancement method based on Retinex multi-level decomposition | |
CN109859138B (en) | Infrared image enhancement method based on human visual characteristics | |
CN114897710A (en) | Video enhancement method under low-illumination environment | |
CN116188339A (en) | Retinex and image fusion-based scotopic vision image enhancement method | |
CN115587945A (en) | High dynamic infrared image detail enhancement method, system and computer storage medium | |
Zeng et al. | High dynamic range infrared image compression and denoising | |
CN112465711A (en) | Degraded image enhancement method for foggy environment | |
CN110852971A (en) | Video defogging method based on dark channel prior and Retinex and computer program product | |
CN111242872A (en) | Real-time RGB image enhancement method based on MSR | |
CN109118458A (en) | A kind of low-luminance color image enchancing method | |
CN109816604B (en) | Self-adaptive clustering infrared image enhancement method |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20220812 |
|
WD01 | Invention patent application deemed withdrawn after publication |