CN107767348B - Single tunnel image rapid enhancement method based on imaging model constraint - Google Patents
Single tunnel image rapid enhancement method based on imaging model constraint Download PDFInfo
- Publication number
- CN107767348B CN107767348B CN201710890293.9A CN201710890293A CN107767348B CN 107767348 B CN107767348 B CN 107767348B CN 201710890293 A CN201710890293 A CN 201710890293A CN 107767348 B CN107767348 B CN 107767348B
- Authority
- CN
- China
- Prior art keywords
- image
- atmospheric light
- tunnel
- local
- imaging model
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000003384 imaging method Methods 0.000 title claims abstract description 28
- 230000005540 biological transmission Effects 0.000 claims abstract description 27
- 238000005286 illumination Methods 0.000 claims abstract description 19
- 238000010586 diagram Methods 0.000 claims description 12
- 238000012937 correction Methods 0.000 claims description 9
- 238000001914 filtration Methods 0.000 claims description 7
- 238000005192 partition Methods 0.000 claims description 7
- 238000012546 transfer Methods 0.000 claims description 6
- 230000005577 local transmission Effects 0.000 claims description 4
- 238000012986 modification Methods 0.000 claims description 4
- 230000004048 modification Effects 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 11
- 238000012545 processing Methods 0.000 abstract description 11
- 230000000694 effects Effects 0.000 abstract description 6
- 238000004088 simulation Methods 0.000 description 3
- 230000000903 blocking effect Effects 0.000 description 2
- 230000002708 enhancing effect Effects 0.000 description 2
- 238000007619 statistical method Methods 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 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
- G06T5/90—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
- 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
- G06T5/90—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
- G06T2207/10004—Still image; Photographic image
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The invention relates to the technical field of image processing, and provides a single tunnel image rapid enhancement method based on imaging model constraint; the parameter estimation and image reconstruction of the imaging model can be completed within seconds, and a tunnel enhanced image with improved brightness and contrast is obtained; the method mainly comprises the following steps of: firstly, atmospheric light estimation; secondly, estimating a transmission map; and thirdly, restoring image calculation. Where the transmission map estimation and illumination map estimation calculations consume a large portion of the computation time. The scheme of the invention avoids two big data statistics and related calculation, and the tunnel image enhancement processing is completed within 3-5 seconds. The image enhancement processing speed is greatly accelerated. The processing effect meets the identification requirement of the tunnel passing condition.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a method for rapidly enhancing a single tunnel image.
Background
The road tunnel environment is closed, and the illumination of the artificial light source is uneven. The resulting image is dark, low contrast, and noisy. The method causes difficulty in quickly and accurately identifying the road passing condition in the tunnel. Therefore, the influence of multicolor haze is effectively weakened, and the contrast of the tunnel image is enhanced as the basis of subsequent image identification. At present, although an image enhancement scheme specially aiming at tunnel image characteristics is not available, a lot of researches on image restoration in foggy days and nighttime foggy days based on imaging model estimation are carried out. Night defogging methods based on illumination estimation, such as suggested by the commander; the weighted entropy method proposed by Dubok Park et al; he Kaiming et al. The methods achieve good defogging effect in actual operation. From the technical idea, these methods are divided into two categories on the basis of imaging models. One estimates model parameters (commander, Dubok Park, etc.) such as scene light source area, illumination intensity, atmospheric light value, transmission diagram, etc. by a statistical method. The method greatly improves the brightness and the contrast of the defogged image, and has clear details and good visual effect. However, because a large amount of statistical operations are required and the processing time is dozens of seconds or more, the method is not suitable for tunnel image enhancement processing with strict real-time requirements.
The other type adds prior knowledge (He Kaiming, etc.) to obtain model additional conditions, and deduces a fixed formula to determine an illumination transmission diagram. Such as dark channel prior. In practice, the method finishes the treatment within a few seconds, and the treatment effect is good. But its application is limited due to the conditions of the dark channel. For tunnel images under artificial light source conditions, the dark channel conditions are apparently not true. Therefore, the method is not good in effect of directly applying to tunnel image enhancement. How to reduce the dependence on the prior knowledge while maintaining the rapidity of the method is an effective improvement method of the method, thereby improving the practicability of the method.
Image defogging and enhancement based on the imaging model comprise three parts of calculation. Firstly, atmospheric light estimation; secondly, estimating a transmission map; and thirdly, restoring image calculation. Some statistical methods also include illumination map estimation. Where the transmission map estimation and illumination map estimation calculations consume a large portion of the computation time. Therefore, reducing the time consumption of transmission map and illumination map estimation greatly reduces the image defogging and enhances the calculation time. On the other hand, more general additional conditions are searched for to solve the model parameter calculation formula, so that the accuracy of a certain degree is guaranteed, and the calculation is fast.
In summary, the invention provides a tunnel image transmission map fast estimation and tunnel image reconstruction algorithm based on imaging model constraints. The parameter estimation and image reconstruction of the imaging model can be completed within seconds. The tunnel enhanced image with improved brightness and contrast is obtained.
Disclosure of Invention
In view of the above, the present invention provides a single tunnel image fast enhancement method based on imaging model constraint; the parameter estimation and image reconstruction of the imaging model can be completed within seconds. The tunnel enhanced image with improved brightness and contrast is obtained.
The invention solves the technical problems by the following technical means:
a single tunnel image rapid enhancement method based on imaging model constraint comprises the following steps:
1) the atmospheric light estimation specifically comprises the following steps:
11) optimizing the whole image by using a quadtree method to determine the global atmospheric light Aglob;
12) Performing uncovered partition on the whole image, and estimating local atmospheric light by using the assumption that the local atmospheric light is a constant
13) Global atmospheric light AglobAnd estimating local atmospheric lightCombined with correction to obtain local atmospheric light correction value
2) The transmission map estimation specifically comprises the following steps:
21) deriving a local transmission map lower limit t according to imaging model constraintsiminA formula;
22) performing non-covering partition on the whole image to locally transmit the lower limit t of the imageiminBased on this, the local transfer diagram t is estimated, following the assumption that the local transfer diagram is constant, by adding the tuning parametersi;
3) Parameter modification, specifically comprising the following steps:
31) obtaining the maximum channel W of the original imagei;
33) with WiFor guidance, estimate t is made for the transmission mapiFiltering is carried out;
4) setting an illumination value to reconstruct a tunnel image, specifically comprising the following steps:
41) carrying out simulation analysis on a large amount of grasped tunnel images to determine the average illumination value L of the images to replace Li;
42) Correcting the corrected local atmospheric lightTransmission map value tiSet illumination map value L and original image valueReconstructing tunnel images by substituting imaging models
Further, in the step 11), the gray scale image of the original image is divided into four equal parts, and the maximum average brightness is obtained; and then, quartering the area where the maximum is located, repeating optimization till 4-5 cycles, wherein the average brightness of the final area is the global atmospheric light estimated value Aglob。
Where D is a local set of pixels. (r)i,gi,bi) Representing points on the corresponding color space; (r)l,gl,bl) Representing the image light source color point.
Further, in the step 13), the local atmospheric light is corrected by combining the following formula:
Further, in the step 22), the local transmission map tiThe estimated formula of (c) is as follows:
wherein, tiIs a partial transmission diagram, timinFor the lower limit of the partial transmission map, θ is an independent variable timinAs a function of (c).
The single tunnel image rapid enhancement method based on the imaging model constraint has the following advantages: according to the scheme of the invention, two big data statistics and related calculation are avoided, so that the tunnel image enhancement processing is completed within 3-5 seconds, the image enhancement processing speed is greatly increased, the brightness and the contrast of the image are improved, and the processing effect meets the tunnel traffic condition identification requirement.
Detailed Description
Fig. 1 shows a flow diagram of a single tunnel image fast enhancement method based on imaging model constraints.
Detailed Description
The present invention will be described in detail below with reference to the accompanying drawings.
Referring to fig. 1, the method for rapidly enhancing a single tunnel image based on imaging model constraint includes the following steps:
1) the atmospheric light estimation specifically comprises the following steps:
11) optimizing the whole image by using a quadtree method to determine the global atmospheric light Aglob(ii) a The method comprises the following specific steps: drawing the gray scale of the original imageDividing equally to obtain the maximum average brightness; and quartering the area where the maximum is located, repeating optimization until the maximum is located for proper times, and obtaining the final average brightness which is the global atmospheric light estimated value Aglob;
12) The whole image is divided into uncovered regions, and the local atmospheric light is estimated by following the assumption that the local atmospheric light is a constant
Where D is a local set of pixels. (r)i,gi,bi) Representing points on the corresponding color space; (r)l,gl,bl) Representing image light source color points;
13) global atmospheric light A is performed by the following formulaglobAnd estimating local atmospheric lightCombined correction to obtain local atmospheric light correction value
2) The transmission map estimation specifically comprises the following steps:
21) deriving a local transmission lower limit t from the imaging model constraintsiminA formula; the pixel value of the normal image should be 0,255]Within the range. The imaging model can be expressed as
Wherein, λ is the number of channels;to restore the channel value of the image at i;is the channel value at i of the pre-image;is the atmospheric light value at i; t is tiIs the transmission map value at i.
The imaging model is therefore constrained by:
thereby deriving a transmission map lower bound timinIs composed of
22) Taking into account blocking artifacts and noise, the weighted transmission map estimation equation is used as follows. Performing non-coverage partition on the whole image to locally transmit the lower limit timinBased on this, the local transfer diagram t is estimated, following the assumption that the local transfer diagram is constant, by adding the tuning parametersi(ii) a Partial transmission map tiThe estimated formula of (c) is as follows:
wherein, tiIs a partial transmission diagram, timinFor the lower limit of the partial transmission map, θIs that the independent variable is timinA certain function of (a); for example,
3) parameter modification, specifically comprising the following steps:
31) obtaining the maximum channel W of the original imagei;
33) with WiFor guidance, estimate t is made for the transmission mapiFiltering is carried out;
due to atmospheric light estimationAnd transmission map estimation value tiThe method is obtained by partitioning under an assumed condition, the partition edge is not smooth, and the blocking effect is obvious; with WiFor guidance, the edges of each partition can be smoothed by filtering processing, so that the image change trend is more accurate.
4) Setting an illumination value to reconstruct a tunnel image, specifically comprising the following steps:
41) carrying out simulation analysis on a large amount of grasped tunnel images to determine the average illumination value L of the images to replace Li(ii) a If the illumination value L is determined statisticallyiThe light source area is determined, the light source intensity is counted, and the estimated transmission graph t is usediCalculating L point by pointi(ii) a The calculation amount is large and time is consumed.
Simulation experiment analysis shows that: for the road tunnel image, a constant value L is taken to replace LiAnd the reconstructed image can achieve good effect.
42) Correcting the corrected local atmospheric lightTransmission map value tiSet illumination map value L and original imageImage valueReconstructing tunnel images by substituting imaging modelsTo make the reconstructed image smoother, W can be usediFor guidance, the reconstructed image is filtered.
Finally, the above embodiments are only for illustrating the technical solutions of the present invention and not for limiting, although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions may be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all of them should be covered in the claims of the present invention.
Claims (5)
1. A single tunnel image rapid enhancement method based on imaging model constraint is characterized in that: the method comprises the following steps:
1) the atmospheric light estimation specifically comprises the following steps:
11) optimizing the whole image by using a quadtree method to determine the global atmospheric light Aglob;
12) Performing uncovered partition on the whole image, and estimating local atmospheric light by using the assumption that the local atmospheric light is a constant
13) Global atmospheric light AglobAnd estimating local atmospheric lightCombined correction to obtain local atmospheric light correction value
2) The transmission map estimation specifically comprises the following steps:
21) according to the imaging modelConstraining, deriving local transmission map lower bound timinA formula;
22) performing non-covering partition on the whole image to locally transmit the lower limit t of the imageiminBased on this, the local transfer diagram t is estimated, following the assumption that the local transfer diagram is constant, by adding the tuning parametersi;
3) Parameter modification, specifically comprising the following steps:
31) obtaining the maximum channel W of the original imagei;
33) with WiFor guidance, estimate t is made for the transmission mapiFiltering is carried out;
4) setting an illumination value to reconstruct a tunnel image, specifically comprising the following steps:
41) simulating and analyzing a plurality of grasped tunnel images to determine an average illumination value L of the images to replace an illumination map Li;
2. The imaging model constraint-based single tunnel image fast enhancement method as claimed in claim 1, characterized in that: in the step 11), the gray level image of the original image is divided into four parts, and the maximum average brightness is obtained; and then, quartering the area where the maximum is located, and repeating the optimization till 4-5 cycles, wherein the average brightness of the final area is theGlobal atmospheric light estimate aglob。
3. The imaging model constraint-based single tunnel image fast enhancement method as claimed in claim 2, characterized in that: in said step 12), the local atmospheric light is estimated by the following formula
Wherein D is a local pixel set; (r)i,gi,bi) Representing points on the corresponding color space; (r)l,gl,bl) Representing the image light source color point.
5. The imaging model constraint-based single tunnel image fast enhancement method of any one of claims 1-4, characterized by: in said step 22), the partial transmission map tiThe estimated formula of (c) is as follows:
wherein, tiIs a partial transmission diagram, timinFor the lower limit of the partial transmission map, θ is an independent variable timinAs a function of (c).
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710890293.9A CN107767348B (en) | 2017-09-27 | 2017-09-27 | Single tunnel image rapid enhancement method based on imaging model constraint |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710890293.9A CN107767348B (en) | 2017-09-27 | 2017-09-27 | Single tunnel image rapid enhancement method based on imaging model constraint |
Publications (2)
Publication Number | Publication Date |
---|---|
CN107767348A CN107767348A (en) | 2018-03-06 |
CN107767348B true CN107767348B (en) | 2021-06-08 |
Family
ID=61266168
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710890293.9A Active CN107767348B (en) | 2017-09-27 | 2017-09-27 | Single tunnel image rapid enhancement method based on imaging model constraint |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107767348B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108805826B (en) * | 2018-05-07 | 2020-12-25 | 珠海全志科技股份有限公司 | Method for improving defogging effect |
CN108600631B (en) * | 2018-05-16 | 2021-03-12 | Oppo广东移动通信有限公司 | Image processing method, image processing device, computer-readable storage medium and electronic equipment |
Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102663694A (en) * | 2012-03-30 | 2012-09-12 | 大连理工大学 | Digital fog effect filter method based on dark primary color channel prior principle |
CN103034983A (en) * | 2013-01-10 | 2013-04-10 | 厦门大学 | Defogging method based on anisotropic filtering |
CN103077504A (en) * | 2013-01-10 | 2013-05-01 | 厦门大学 | Image haze removal method on basis of self-adaptive illumination calculation |
CN103198459A (en) * | 2013-04-10 | 2013-07-10 | 成都国腾电子技术股份有限公司 | Haze image rapid haze removal method |
CN103500440A (en) * | 2013-09-27 | 2014-01-08 | 中国人民解放军空军工程大学 | Method for eliminating cloud and haze of atmospheric degraded image |
US8755628B2 (en) * | 2012-09-10 | 2014-06-17 | Google Inc. | Image de-hazing by solving transmission value |
CN104063853A (en) * | 2014-07-07 | 2014-09-24 | 南京通用电器有限公司 | Method for improving traffic video image definition based on dark channel technology |
CN106530257A (en) * | 2016-11-22 | 2017-03-22 | 重庆邮电大学 | Remote sensing image de-fogging method based on dark channel prior model |
-
2017
- 2017-09-27 CN CN201710890293.9A patent/CN107767348B/en active Active
Patent Citations (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102663694A (en) * | 2012-03-30 | 2012-09-12 | 大连理工大学 | Digital fog effect filter method based on dark primary color channel prior principle |
US8755628B2 (en) * | 2012-09-10 | 2014-06-17 | Google Inc. | Image de-hazing by solving transmission value |
CN103034983A (en) * | 2013-01-10 | 2013-04-10 | 厦门大学 | Defogging method based on anisotropic filtering |
CN103077504A (en) * | 2013-01-10 | 2013-05-01 | 厦门大学 | Image haze removal method on basis of self-adaptive illumination calculation |
CN103198459A (en) * | 2013-04-10 | 2013-07-10 | 成都国腾电子技术股份有限公司 | Haze image rapid haze removal method |
CN103500440A (en) * | 2013-09-27 | 2014-01-08 | 中国人民解放军空军工程大学 | Method for eliminating cloud and haze of atmospheric degraded image |
CN104063853A (en) * | 2014-07-07 | 2014-09-24 | 南京通用电器有限公司 | Method for improving traffic video image definition based on dark channel technology |
CN106530257A (en) * | 2016-11-22 | 2017-03-22 | 重庆邮电大学 | Remote sensing image de-fogging method based on dark channel prior model |
Non-Patent Citations (5)
Title |
---|
NIGHTTIME IMAGE DEHAZING WITH LOCAL ATMOSPHERIC LIGHT AND WEIGHTED ENTROPY;Dubok Park,David K. Han,Hanseok Ko;《IEEE International Conference on Image Processing》;20161231;第2261-2265页 * |
SINGLE IMAGE DEHAZING WITH IMAGE ENTROPY AND INFORMATION FIDELITY;Dubok Park,Hyungjo Park,David K. Han,Hanseok Ko;《IEEE International Conference on Image Processing》;20141231;第4037-4041页 * |
基于光照估计的夜间图像去雾;方帅,赵育坤,李心科,刘永进,揭斐然;《电子学报》;20161130;第2569-2575页 * |
基于自适应透射率修复的去雾算法研究;兰军明,闭应洲,邱骏达,韩冬越;《 广西师范学院学报:自然科学版》;20160331;第64-69页 * |
基于超像素分割的图像去雾算法;郑良缘,王平,高颖慧;《重庆理工大学学报(自然科学)》;20150630;第100-106页 * |
Also Published As
Publication number | Publication date |
---|---|
CN107767348A (en) | 2018-03-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106157267B (en) | Image defogging transmissivity optimization method based on dark channel prior | |
CN108734670B (en) | Method for restoring single night weak-illumination haze image | |
Gao et al. | Sand-dust image restoration based on reversing the blue channel prior | |
CN105046663B (en) | A kind of adaptive enhancement method of low-illumination image for simulating human visual perception | |
CN109255758B (en) | Image enhancement method based on all 1 x 1 convolution neural network | |
CN107292830B (en) | Low-illumination image enhancement and evaluation method | |
CN103218778A (en) | Image and video processing method and device | |
CN104867121B (en) | Image Quick demisting method based on dark primary priori and Retinex theories | |
CN105303532A (en) | Wavelet domain Retinex image defogging method | |
KR101361644B1 (en) | Image processing device and image defogging method | |
CN108133462B (en) | Single image restoration method based on gradient field region segmentation | |
CN110288550A (en) | The single image defogging method of confrontation network is generated based on priori knowledge guiding conditions | |
CN104331867B (en) | The method, device and mobile terminal of image defogging | |
CN104091307A (en) | Frog day image rapid restoration method based on feedback mean value filtering | |
CN107767348B (en) | Single tunnel image rapid enhancement method based on imaging model constraint | |
CN111598814B (en) | Single image defogging method based on extreme scattering channel | |
CN110807744A (en) | Image defogging method based on convolutional neural network | |
CN102646267A (en) | Degraded image restoration method and system | |
CN110349113B (en) | Adaptive image defogging method based on dark primary color priori improvement | |
CN108629750A (en) | A kind of night defogging method, terminal device and storage medium | |
CN104168402A (en) | Method and device for video frame image defogging | |
CN108765337B (en) | Single color image defogging processing method based on dark channel prior and non-local MTV model | |
Yang et al. | A variation-based nighttime image dehazing flow with a physically valid illumination estimator and a luminance-guided coloring model | |
CN111091580B (en) | Stumpage image segmentation method based on improved ResNet-UNet network | |
CN117611501A (en) | Low-illumination image enhancement method, device, equipment and readable storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right |
Effective date of registration: 20230208 Address after: 213399 room 5025, building B, 218 Hongkou Road, Kunlun Street, Liyang City, Changzhou City, Jiangsu Province Patentee after: Liyang Smart City Research Institute of Chongqing University Address before: 400044 No. 174 Sha Jie street, Shapingba District, Chongqing Patentee before: Chongqing University |
|
TR01 | Transfer of patent right |