CN104915665A - Image defogging method and license plate image identification method based on the method - Google Patents
Image defogging method and license plate image identification method based on the method Download PDFInfo
- Publication number
- CN104915665A CN104915665A CN201510292684.1A CN201510292684A CN104915665A CN 104915665 A CN104915665 A CN 104915665A CN 201510292684 A CN201510292684 A CN 201510292684A CN 104915665 A CN104915665 A CN 104915665A
- Authority
- CN
- China
- Prior art keywords
- image
- license plate
- defogging
- optimization
- matrix
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
- G06V10/12—Details of acquisition arrangements; Constructional details thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Abstract
The invention discloses an image defogging method and a license plate image identification method based on the method by combining a dark channel optimization defogging method and a two-dimensional hidden Markov model. The image defogging method comprises steps of getting a dark channel for a foggy original image obtained through photographing to obtain an original transmission image, performing optimization on the original transmission image to obtain an optimization transmission image, obtaining a corresponding array, performing second optimization calculation on the original transmission image to obtain a final transmission image, and utilizing the final transmission image to obtain a defogging image. The invention enables the defogging result to be enhanced at the blooming effect at the depth discontinuous part, utilizes the condition transfer probability in the image recognition to perform state switching, reflects the dependence relation in two dimensions of the plane and identifies the English characters and the Arabic numbers. For the font size, font inclination and fouling, the image defogging method is strong in resisting noise.
Description
Technical field
The present invention relates to image procossing and field of target recognition, particularly relate to a kind of image defogging method capable and the license plate image recognition methods based on the method.
Background technology
Due to the impact of atmospheric aerosol (suspended particle, mist, haze etc.), cause the image quality decrease of taking, the contrast of image declines and causes color distortion in addition, recovers the contrast and the true colors that there are mist image, in photography and computer vision, has important application.First, image mist elimination can increase the visibility of scene and revise misalignment; Secondly, most image processing algorithm all based on the radiation value that the pixel value of input picture is scene, when color generation deviation or contrast decline, can have a negative impact to Processing Algorithm usually; Again, the depth information that image mist elimination algorithm obtains, provides extra Data support by giving many computer vision algorithms make.
The defogging method capable of single image has had remarkable progress in recent years, processing mode on a priori assumption basis, statistics shows that most of natural scene image meets this hypothesis, in conjunction with to the physical model having image, dark method is widely used in single image mist elimination, but it exists the problem obviously declined from viewpoint mist elimination effect remotely, occur that figure is dizzy at the discontinuous place of the degree of depth.
Summary of the invention
In order to overcome above-mentioned the deficiencies in the prior art, the invention provides a kind of image defogging method capable and the license plate image recognition methods based on the method, the present invention eliminates existing method process image and to swoon effect at the figure that the discontinuous place of the degree of depth occurs, remotely high-visible in viewpoint by the image of this method process.
To achieve these goals, concrete steps of the present invention are as follows:
Step one: the mist original image that has obtained taking pictures asks for dark J
dark, calculate skylight brightness S, obtain initial transmission figure
Step 2: at weights K=10
-4condition under, use formula (1) to initial transmission figure
be optimized, be optimized transmission plot t ', carries out expansion process, the matrix P that be optimized weighted after carrying out edge extracting to optimization transmission plot t ';
(wherein I is the pixel value of original image, and J is the true radiation value expecting the scene obtained, and A is skylight brightness, and t is the transmissivity of medium);
Step 3: the initial transmission figure that the weighted matrix P of the optimization using formula (2) to be obtained by step 2 is obtained step one
again optimize calculating, obtain final transmission plot t,
(K is weights, and P is weighting matrix, and L is that Laplce scratches figure matrix (Matting Laplacian matrix), and t is final transmission plot,
for initial transmission figure);
Step 4: utilize the result that step 3 obtains, calculates mist elimination image J by formula (1) to final transmission plot t.
Present invention also offers a kind of Hidden Markov license plate image recognition methods based on optimizing dark channel image defogging method capable, comprising:
Step 5: carry out License Plate to the mist elimination image that step 5 obtains, by horizontal projection and vertical projection positioning licence plate, obtains license plate image;
Step 6: adopt two-dimentional Hidden Markov Model (HMM) to carry out character recognition to license plate image, obtain character identification result.
Compared with prior art, the present invention can be optimized having mist original image, has extraordinary mist elimination effect, obtains mist elimination image comparatively clearly, and the figure that there is not the viewpoint discontinuous appearance of the fuzzy and degree of depth remotely in mist elimination image swoons.
License plate image recognition methods of the present invention adopts two-dimentional Hidden Markov Model (HMM) to carry out character recognition to license plate image, wherein each state is all by the model-composing of a subordinate Hidden Markov, major state is horizontal direction, be vertical direction from state, the optimization dark method processed under being applicable to depth of field fuzzy enviroment, image recognition under the bad weather circumstances such as solution haze weather has a significant effect, this algorithm is applicable to the Car license recognition in greasy weather situation, kinds of characters size, character are tilted, the situation such as stained, noise resisting ability is strong.
Accompanying drawing explanation
Fig. 1 be take pictures obtain have mist original image;
Fig. 2 is the image that Fig. 1 obtains after defogging method capable process of the present invention;
Fig. 3 be in Fig. 2 car plate through partial enlargement and the recognition result obtained after identifying;
Fig. 4 is the workflow diagram of license plate image recognition methods of the present invention.
Embodiment
Below in conjunction with embodiment, the present invention will be further described.
See Fig. 4, defogging method capable of the present invention comprises the following steps:
Step one: the mist original image that has obtained taking pictures asks for dark J
dark, calculate skylight brightness S, obtain initial transmission figure
Step 2: at weights K=10
-4condition under, use formula (1) to initial transmission figure
be optimized, be optimized transmission plot t ', carries out expansion process, the matrix P that be optimized weighted after carrying out edge extracting to optimization transmission plot t ';
(wherein I is the pixel value of original image, and J is the true radiation value expecting the scene obtained, and A is skylight brightness, and t is the transmissivity of medium);
Step 3: the initial transmission figure that the weighted matrix P of the optimization using formula (2) to be obtained by step 2 is obtained step one
again optimize calculating, obtain final transmission plot t,
(K is weights, and P is weighting matrix, and L is that Laplce scratches figure matrix (Matting Laplacian matrix), and t is final transmission plot,
for initial transmission figure);
Step 4: utilize the result that step 3 obtains, calculates mist elimination image J by formula (1) to final transmission plot t.
License plate image recognition methods provided by the invention is further comprising the steps of:
Step 5: carry out License Plate to the mist elimination image that step 5 obtains, by horizontal projection and vertical projection positioning licence plate, obtains license plate image;
Step 6: adopt two-dimentional Hidden Markov Model (HMM) to carry out character recognition to license plate image, obtain character identification result.
See Fig. 1, Fig. 1 be take pictures obtain have mist original image, the non-constant of picture clarity, see Fig. 2, the image that Fig. 2 obtains after being through defogging method capable process of the present invention, as can be seen from Figure 2 mist elimination effect of the present invention is very good, and the figure that there is not the viewpoint discontinuous appearance of the fuzzy and degree of depth remotely in mist elimination image swoons.
Be the character identification result obtained see Fig. 3, Fig. 3, swooned by the figure that there is not the viewpoint discontinuous appearance of the fuzzy and degree of depth remotely in the mist elimination image of process of the present invention, reduce the degree of difficulty that the later stage carries out greasy weather Car license recognition.
Claims (2)
1. an image defogging method capable, is characterized in that, concrete steps are as follows:
Step one: the mist original image that has obtained taking pictures asks for dark J
dark, calculate skylight brightness S, obtain initial transmission figure
Step 2: at weights K=10
-4condition under, use formula (1) to initial transmission figure
be optimized, be optimized transmission plot t ', carries out expansion process, the matrix P that be optimized weighted after carrying out edge extracting to optimization transmission plot t ';
(wherein I is the pixel value of original image, and J is the true radiation value expecting the scene obtained, and A is skylight brightness, and t is the transmissivity of medium);
Step 3: the initial transmission figure that the weighted matrix P of the optimization using formula (2) to be obtained by step 2 is obtained step one
again optimize calculating, obtain final transmission plot t,
(K is weights, and P is weighting matrix, and L is that Laplce scratches figure matrix (Matting Laplacian matrix), and t is final transmission plot,
for initial transmission figure);
Step 4: utilize the result that step 3 obtains, calculates mist elimination image J by formula (1) to final transmission plot t.
2., based on a license plate image recognition methods for method described in claim 1, it is characterized in that, comprising:
Step 5: carry out License Plate to the mist elimination image that step 5 obtains, by horizontal projection and vertical projection positioning licence plate, obtains license plate image;
Step 6: adopt two-dimentional Hidden Markov Model (HMM) to carry out character recognition to license plate image, obtain character identification result.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510292684.1A CN104915665A (en) | 2015-06-01 | 2015-06-01 | Image defogging method and license plate image identification method based on the method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201510292684.1A CN104915665A (en) | 2015-06-01 | 2015-06-01 | Image defogging method and license plate image identification method based on the method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN104915665A true CN104915665A (en) | 2015-09-16 |
Family
ID=54084715
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201510292684.1A Pending CN104915665A (en) | 2015-06-01 | 2015-06-01 | Image defogging method and license plate image identification method based on the method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN104915665A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109492554A (en) * | 2018-10-25 | 2019-03-19 | 烟台市奥境数字科技有限公司 | A kind of intelligent traffic monitoring image wears mist recognition methods |
CN112598002A (en) * | 2020-12-07 | 2021-04-02 | 南京航空航天大学 | License plate recognition method under influence of fog and noise |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8290294B2 (en) * | 2008-09-16 | 2012-10-16 | Microsoft Corporation | Dehazing an image using a three-dimensional reference model |
US20130071043A1 (en) * | 2011-09-08 | 2013-03-21 | Fujitsu Limited | Image defogging method and system |
CN103020902A (en) * | 2011-09-20 | 2013-04-03 | 富士通株式会社 | Exposure enhancement method and exposure enhancement device for mist elimination images |
CN104200207A (en) * | 2014-09-16 | 2014-12-10 | 宁波熵联信息技术有限公司 | License plate recognition method based on Hidden Markov models |
-
2015
- 2015-06-01 CN CN201510292684.1A patent/CN104915665A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8290294B2 (en) * | 2008-09-16 | 2012-10-16 | Microsoft Corporation | Dehazing an image using a three-dimensional reference model |
US20130071043A1 (en) * | 2011-09-08 | 2013-03-21 | Fujitsu Limited | Image defogging method and system |
CN103020902A (en) * | 2011-09-20 | 2013-04-03 | 富士通株式会社 | Exposure enhancement method and exposure enhancement device for mist elimination images |
CN104200207A (en) * | 2014-09-16 | 2014-12-10 | 宁波熵联信息技术有限公司 | License plate recognition method based on Hidden Markov models |
Non-Patent Citations (2)
Title |
---|
刘楠等: "基于加权暗通道的图像去雾方法", 《光电子学报》 * |
陈友仁等: ""基于隐马尔可夫模型的车牌自动识别技术"", 《红外与激光工程》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109492554A (en) * | 2018-10-25 | 2019-03-19 | 烟台市奥境数字科技有限公司 | A kind of intelligent traffic monitoring image wears mist recognition methods |
CN112598002A (en) * | 2020-12-07 | 2021-04-02 | 南京航空航天大学 | License plate recognition method under influence of fog and noise |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110570371B (en) | Image defogging method based on multi-scale residual error learning | |
JP6789402B2 (en) | Method of determining the appearance of an object in an image, equipment, equipment and storage medium | |
CN103218778B (en) | The disposal route of a kind of image and video and device | |
CN102831591B (en) | Gaussian filter-based real-time defogging method for single image | |
CN110544213B (en) | Image defogging method based on global and local feature fusion | |
CN107767354A (en) | A kind of image defogging algorithm based on dark primary priori | |
CN104318524A (en) | Method, device and system for image enhancement based on YCbCr color space | |
CN109509156B (en) | Image defogging processing method based on generation countermeasure model | |
CN103559693A (en) | Image local structure self-adaption recovery method based on non-continuity indicator | |
CN110675340A (en) | Single image defogging method and medium based on improved non-local prior | |
Ding et al. | Single-image haze removal using the mean vector L2-norm of RGB image sample window | |
CN110807738B (en) | Blurred image non-blind restoration method based on edge image block sharpening | |
CN104281999A (en) | Single image defogging method based on structural information | |
CN103034983A (en) | Defogging method based on anisotropic filtering | |
CN106023108A (en) | Image defogging algorithm based on boundary constraint and context regularization | |
CN116038147A (en) | Laser metal cutting machine and control method thereof | |
CN105989583A (en) | Image defogging method | |
CN105023246A (en) | Image enhancement method based on contrast and structural similarity | |
CN104915665A (en) | Image defogging method and license plate image identification method based on the method | |
CN104065958A (en) | LED display screen brightness collecting system | |
CN111414938B (en) | Target detection method for bubbles in plate heat exchanger | |
CN112465720A (en) | Image defogging method and device based on image sky segmentation and storage medium | |
CN110738624B (en) | Area-adaptive image defogging system and method | |
CN110399886B (en) | Screen image JND model construction method | |
CN109191405B (en) | Aerial image defogging algorithm based on transmittance global estimation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150916 |
|
RJ01 | Rejection of invention patent application after publication |