CN106355180B - A kind of license plate locating method combined based on color with edge feature - Google Patents
A kind of license plate locating method combined based on color with edge feature Download PDFInfo
- Publication number
- CN106355180B CN106355180B CN201610807379.6A CN201610807379A CN106355180B CN 106355180 B CN106355180 B CN 106355180B CN 201610807379 A CN201610807379 A CN 201610807379A CN 106355180 B CN106355180 B CN 106355180B
- Authority
- CN
- China
- Prior art keywords
- license plate
- candidate
- image
- svm
- color
- 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
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
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/56—Extraction of image or video features relating to colour
-
- 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
Abstract
The invention discloses a kind of license plate locating methods combined based on color with edge feature, belong to field of intelligent transportation technology.This method uses the color characteristic of license plate to carry out License Plate to input picture first, then carries out license plate judgement using SVM model, obtains license plate;If not orienting license plate, License Plate is carried out to input picture using the edge feature of license plate, license plate judgement is then carried out using SVM model, obtains license plate.For the license plate image of background complexity, the present invention can fast and effeciently orient license plate.
Description
Technical field
The present invention relates to a kind of license plate locating methods combined based on color with edge feature, belong to intelligent transport technology
Field.
Background technique
With the fast development of Modern Traffic, the modern management of vehicle is increasingly tended to automate, vehicle is examined automatically
It surveys, the demand of identifying system increasingly increases, such as high speed crossing charging aperture, cell automatic parking lot management etc..It is existing in order to adapt to
For the development of traffic, intelligentized traffic control system is come into being, and one of important link is namely based on the identification of license plate
System.License plate recognition technology is that a very important research topic, license plate recognition technology are main in modern intelligent transportation field
Including three parts such as License Plate, License Plate Character Segmentation and Recognition of License Plate Characters, and during License Plate is Car license recognition
One link of most critical, directly affects the performance of Vehicle License Plate Recognition System.
For the license plate image of background complexity, how correctly, License Plate is quickly and efficiently carried out, is that current license plate is known
One of most important task in other system.Currently, license plate locating method is broadly divided into two classes, i.e. the license plate based on gray level image is fixed
Position method and the license plate locating method based on color image.License plate locating method based on gray level image mainly uses edge detection
Operator carries out edge detection to gray level image, to obtain candidate license plate.Common edge detection operator has the inspection of the edge Roberts
Measuring and calculating, Sobel edge detection operator and Prewitt boundary operator etc..Experiment shows that Sobel edge detection operator can not only
Preferably prominent license plate edge feature, and fast speed.Based on the localization method of Sobel edge detection operator, detailed process
As indicated with 1, this method carries out Gaussian Blur to the color image of input first and filters interference noise figure, and to the image of denoising into
Row gray processing;Then the vertical edge of gray level image is extracted using Sobel edge detection operator, while two are carried out to edge image
Value processing, and using all profiles of closed operation extraction image;Finally according to the area of license plate, length-width ratio etc. to profile most
Small boundary rectangle is screened, to extract candidate license plate.But the license plate locating method based on Sobel operator has ignored
The colouring information of color image interlocks although many license plates can be oriented fast and effeciently in license plate image vertical edge
In the case where, this method can not orient license plate.Meanwhile this method can only cursorily orient license plate.
Compared with color image, gray level image has lost colouring information, so color image is more conducive to License Plate, and
License plate can be accurately positioned out when complex background, different illumination.There is common color model in Digital Image Processing
RGB model and HSV model, 3 components Rs, G, B of RGB model change with the change of brightness, and HSV model coloration
H, 3 saturation degree S, brightness V representation in components, wherein H, S contain the colour information of image, are not influenced by brightness, and V component
Luminance information is only contained, 3 components are mutually indepedent.Therefore, HSV model is mainly used using the license plate locating method of color.
License plate locating method based on HSV space is as shown in Fig. 2, this method first converts the RGB image of input in the figure of HSV space
Picture;Then H, S, V threshold range that setting is respectively adopted carry out binaryzation to the image of HSV space, and using closed operation and mention
Take all profiles;Finally according to the tilt angle of license plate, length-width ratio and edge divide error to the minimum circumscribed rectangle of profile into
Row screening, to obtain license plate.Compared with the license plate locating method based on Sobel operator, the localization method benefit based on HSV space
With colouring information, the success rate of License Plate is improved.But for license plate color is identical as body color, illumination is insufficient or
Situations such as exposure, this method can not efficiently locate out license plate.Meanwhile the localization method is mainly according to artificial experience to candidate vehicle
Board is screened, and robustness is poor.
Summary of the invention
The technical problem to be solved by the present invention is to overcome the deficiencies of existing technologies, provide a kind of special based on color and edge
The license plate locating method combined is levied, can fast and effeciently orient license plate in the license plate image of background complexity.
In order to solve the above technical problems, the present invention provides a kind of License Plate side combined based on color with edge feature
Method, comprising the following steps:
1) using license plate color characteristic to input picture carry out License Plate, then using SVM license plate discrimination model into
Board of driving a vehicle judges, obtains license plate;
If 2) be greater than maxPlate by the license plate quantity that color characteristic is oriented, do not need to carry out Sobel twice
Otherwise positioning is carried out License Plate to input picture using the edge feature of license plate, is then carried out using SVM license plate discrimination model
License plate judgement, obtains license plate;The maxPlate indicates the number of license plate in a width license plate image.
Above-mentioned state carries out License Plate to input picture using color characteristic, the specific steps are as follows:
HSV space 1-1) is converted from rgb space by the color space for inputting license plate image, while carrying out histogram equalization
Change processing;
Blue matching template and yellow matching template 1-2) is respectively adopted, binaryzation is carried out to pretreated license plate image
Processing;
Closed operation 1-3) is carried out to the license plate image of binaryzation, then extracts all profiles of the image;
Minimum circumscribed rectangle 1-4) is taken to each profile, then according to the tilt angle of license plate and length-width ratio primary filtration one
A little underproof candidate license plates, are finally standardized the size of license plate;
1-5) carried out using trained SVM license plate discrimination model to by the filtered remaining candidate license plate of step 1-4)
Judgement obtains license plate.
Edge feature above-mentioned using license plate carries out License Plate to input picture, the specific steps are as follows:
The processed filter interference noise of Gaussian mode gelatinization 2-1) is carried out to the license plate image of input, ash then is carried out to the image
Degreeization;
2-2) horizontal and vertical edge is extracted to the image of gray processing by Sobel operator respectively, obtains the edge of license plate
Image;
Binaryzation, closed operation processing successively 2-3) are carried out to step 2-2) the license plate edge image obtained, then extract institute
There is profile, repairing treatment finally carried out to the chain rupture part in profile, while edge image is obtained using Sobel operator again,
To extract all profiles;
Minimum circumscribed rectangle 2-4) is taken to each profile, while according to the tilt angle of license plate and length-width ratio primary filtration one
A little underproof candidate license plates, are then standardized the size of remaining candidate license plate;
2-5) carried out using trained SVM license plate discrimination model to by the filtered remaining candidate license plate of step 2-4)
Judgement obtains license plate.
It is above-mentioned to be referred to according to some underproof candidate license plates of tilt angle and length-width ratio primary filtration of license plate, if
The absolute value of license plate sloped angle is less than 30 degree, and the length-width ratio of license plate then retains the candidate license plate, otherwise give up between 2-4
Abandon the candidate license plate.
The acquisition of SVM license plate discrimination model above-mentioned, the specific steps are as follows:
A large amount of candidate license plate 3-1) is generated, color characteristic and edge feature is respectively adopted, input license plate image is determined
Position, obtains a large amount of candidate license plate;
3-2) label for candidate license plate;
License plate training set 3-3) is constructed, candidate license plate includes the real license plate of two classes and non-license plate, respectively from real license plate figure
A certain number of pictures are extracted in piece collection and non-license plate pictures and constitute license plate training set, and remaining picture constitutes license plate test
Collection;
3-4) training SVM license plate discrimination model, is trained SVM model by license plate training set, obtains SVM license plate and sentences
Other model, while SVM license plate discrimination model is verified using license plate test set.
Aforementioned step 3-2) in, the method for using successive iteration automatic labeling label is labelled for each candidate license plate, tool
Body is shown in steps are as follows:
The candidate license plate of a certain number of unlabelleds 3-2-1) is chosen, while these candidate license plates are manually labelled
Label;
3-2-2) candidate license plate labeled is trained using SVM model, a SVM license plate is obtained and differentiates
Model;
The candidate license plate for 3-2-3) choosing a certain number of unlabelleds again, using trained SVM license plate discrimination model
To the candidate license plate automatic labeling label of these unlabelleds, while manual confirmation labels wrong candidate license plate;
3-2-4) candidate license plate correctly labeled is combined, then carries out step 3-2-2), until
All candidate license plates all correctly post label.
Aforementioned step 3-3) in, the picture that 70% is extracted from real license plate pictures and non-license plate pictures constitutes vehicle
Board training set.
MaxPlate value above-mentioned takes 1.
Size above-mentioned to license plate, which is standardized, to be referred to, the size of pick-up board is 136*36.
Advantageous effects of the invention:
The present invention is directed to the license plate image of background complexity, can fast and effeciently orient license plate.
Detailed description of the invention
Fig. 1 is the license plate locating method flow chart based on Sobel boundary operator;
Fig. 2 is the license plate locating method flow chart based on HSV space;
Fig. 3 is the training of SVM license plate discrimination model and the process for judging license plate;
Fig. 4 is the flow chart of license plate locating method of the invention.
Specific embodiment
The invention will be further described below in conjunction with the accompanying drawings.Following embodiment is only used for clearly illustrating the present invention
Technical solution, and not intended to limit the protection scope of the present invention.
The license plate locating method combined based on color with edge feature of the invention, as shown in figure 4, including following step
It is rapid:
1, using the color characteristic of license plate to input picture carry out License Plate, then using SVM license plate discrimination model into
Board of driving a vehicle judges, obtains license plate.License Plate is carried out to input picture using color characteristic, the specific steps are as follows:
HSV space 1-1) is converted from rgb space by the color space for inputting license plate image, while carrying out histogram equalization
Change processing;
Blue matching template and yellow matching template 1-2) is respectively adopted, binaryzation is carried out to pretreated license plate image
Processing;Such as: for each pixel in image, if the value of H is fallen between 95-140 and S value and V value respectively fall in 70-
Between 255 and 60-255, then white pixel is set by the pixel, is otherwise black picture element.Table 1 is respectively blue and yellow
H, S, V value range with template.
The matching template of 1 yellow of table and blue
H_min | H_max | S_min | S_max | V_min | V_max | |
Blue matching template | 95 | 140 | 70 | 255 | 60 | 255 |
Yellow matching template | 12 | 40 | 70 | 255 | 60 | 255 |
Closed operation 1-3) is carried out to the license plate image of binaryzation, then extracts all profiles of the image;
Minimum circumscribed rectangle 1-4) is taken to each profile, it is then big according to the tilt angle of license plate and length-width ratio primary filtration
The underproof candidate license plate in part, is finally standardized the size of license plate;The size of general China's license plate is
440mm*140mm, length-width ratio 3.14, and the tilt angle of license plate will not be very big in practical application, generally arrives in -30 degree
Between 30 degree.If the absolute value of license plate sloped angle is less than 30 degree, and the length-width ratio of license plate then retains the time between 2-4
License plate is selected, the candidate license plate is otherwise given up.Therefore, which can filter 80% underproof candidate license plate.
1-5) candidate license plate remaining after step 1-4) is judged using trained SVM license plate discrimination model,
Obtain license plate.
SVM license plate discrimination model requires the size of every license plate identical, otherwise can not license plate be learnt and be differentiated.Root
The mean breadth and average height of a few thousand sheets license plates according to statistics, the size of pick-up board of the present invention are 136*36.
If 2, being greater than maxPlate by the license plate quantity that color characteristic is oriented, do not need to carry out Sobel twice
Otherwise positioning is carried out License Plate to input picture using the edge feature of license plate, is then carried out using SVM license plate discrimination model
License plate judgement, obtains license plate.MaxPlate indicates the number of license plate in a width license plate image, and the value is 1 in the present invention.
License Plate is carried out to input picture using the edge feature of license plate, the specific steps are as follows:
The processed filter interference noise of Gaussian mode gelatinization 2-1) is carried out to the license plate image of input, ash then is carried out to the image
Degreeization;
2-2) horizontal and vertical edge is extracted to the image of gray processing by Sobel operator respectively, obtains the edge of license plate
Image;
Binaryzation, closed operation processing successively 2-3) are carried out to step 2-2) the license plate edge image obtained, then extract institute
There is profile, repairing treatment finally is carried out to the chain rupture part in profile, while more accurately edge is obtained using Sobel operator
Image, to extract all profiles;
Minimum circumscribed rectangle 2-4) is taken to each profile, while according to the length-width ratio of license plate and tilt angle primary filtration one
A little underproof candidate license plates, are then standardized the size of remaining candidate license plate;It is identical as step 1-4) to filter principle;
2-5) candidate license plate remaining after step 2-4) is sentenced using trained SVM license plate discrimination model
It is disconnected, obtain license plate.
In the step, license plate locating method of the hardware realization based on Sobel boundary operator can be used.
Step 1-5) and step 2-5) in SVM license plate discrimination model acquisition, as shown in Figure 3, the specific steps are as follows:
3-1) generate a large amount of candidate license plate.Color characteristic and edge feature is respectively adopted to determine input license plate image
Position, obtains a large amount of candidate license plate.Candidate license plate includes real license plate and non-two class of license plate.
3-2) label for candidate license plate.It is manually a very big workload for the labelling of each candidate license plate, this
A kind of method for using successive iteration automatic labeling label is invented to label for each candidate license plate.This method is first to having pasted
The candidate license plate collection training of label obtains a SVM license plate discrimination model, then using the model to the candidate vehicle of unlabelled
Board collection carries out automatic labeling label.The process is constantly repeated, until all candidate license plates are all correctly labeled, specific step is as follows
It is shown:
3-2-1) the candidate license plate of selected part unlabelled, while manually being labelled to these candidate license plates;
3-2-2) candidate license plate labeled is trained using SVM model, a SVM license plate is obtained and differentiates
Model;
3-2-3) the candidate license plate of selected part unlabelled, using trained SVM license plate discrimination model to these times
License plate automatic labeling label are selected, while manual confirmation labels wrong candidate license plate;
3-2-4) candidate license plate correctly labeled is combined, then carries out step 3-2-2), until
All candidate license plates all correctly post label.
3-3) construct license plate training set.Candidate license plate includes the real license plate of two classes and non-license plate, respectively from real license plate figure
Extract a certain number of pictures in piece collection and non-license plate pictures and constitute license plate training sets, i.e., respectively from real license plate picture with it is non-
The picture that 70% is obtained in license plate picture constitutes license plate training set, and remaining picture constitutes license plate test set, tests training SVM vehicle
Board discrimination model.
3-4) training SVM license plate discrimination model.SVM model is trained by license plate training set, SVM license plate is obtained and sentences
Other model, while SVM license plate discrimination model is verified using license plate test set.
The present invention differentiates candidate license plate using SVM model, and SVM model is machine learning algorithm.Therefore, may be used
To be differentiated using other machine learning algorithms such as neural network (BP) to candidate license plate.
The above is only a preferred embodiment of the present invention, it is noted that for the ordinary skill people of the art
For member, without departing from the technical principles of the invention, several improvement and deformations can also be made, these improvement and deformations
Also it should be regarded as protection scope of the present invention.
Claims (7)
1. a kind of license plate locating method combined based on color with edge feature, which comprises the following steps:
1) License Plate is carried out to input picture using the color characteristic of license plate, vehicle is then carried out using SVM license plate discrimination model
Board judgement, obtains license plate;
If 2) be greater than maxPlate by the license plate quantity that color characteristic is oriented, do not need to carry out Sobel twice and positions,
Otherwise License Plate is carried out to input picture using the edge feature of license plate, license plate is then carried out using SVM license plate discrimination model
Judgement obtains license plate;The maxPlate indicates the number of license plate in a width license plate image;
The acquisition of the SVM license plate discrimination model, the specific steps are as follows:
A large amount of candidate license plate 3-1) is generated, color characteristic and edge feature is respectively adopted, input license plate image is positioned,
Obtain a large amount of candidate license plate;
3-2) label for candidate license plate;The method for using successive iteration automatic labeling label is labelled for each candidate license plate, tool
Body is shown in steps are as follows:
The candidate license plate of a certain number of unlabelleds 3-2-1) is chosen, while manually being labelled to these candidate license plates;
3-2-2) candidate license plate labeled is trained using SVM model, a SVM license plate is obtained and differentiates mould
Type;
The candidate license plate for 3-2-3) choosing a certain number of unlabelleds again, using trained SVM license plate discrimination model to this
The candidate license plate automatic labeling label of a little unlabelleds, while manual confirmation labels wrong candidate license plate;
3-2-4) candidate license plate correctly labeled is combined, then carries out step 3-2-2), until all
Candidate license plate all correctly posts label;
License plate training set 3-3) is constructed, candidate license plate includes the real license plate of two classes and non-license plate, respectively from real license plate pictures
License plate training set is constituted with a certain number of pictures are extracted in non-license plate pictures, remaining picture constitutes license plate test set;
3-4) training SVM license plate discrimination model, is trained SVM model by license plate training set, obtains SVM license plate and differentiates mould
Type, while SVM license plate discrimination model is verified using license plate test set.
2. a kind of license plate locating method combined based on color with edge feature according to claim 1, feature are existed
In the color characteristic using license plate carries out License Plate to input picture, the specific steps are as follows:
HSV space 1-1) is converted from rgb space by the color space for inputting license plate image, while being carried out at histogram equalization
Reason;
Blue matching template and yellow matching template 1-2) is respectively adopted, binary conversion treatment is carried out to pretreated license plate image;
Closed operation 1-3) is carried out to the license plate image of binaryzation, then extracts all profiles of the image;
Minimum circumscribed rectangle 1-4) is taken to each profile, it is then some not according to the tilt angle of license plate and length-width ratio primary filtration
Qualified candidate license plate, is finally standardized the size of license plate;
1-5) sentenced using trained SVM license plate discrimination model to by the filtered remaining candidate license plate of step 1-4)
It is disconnected, obtain license plate.
3. a kind of license plate locating method combined based on color with edge feature according to claim 1, feature are existed
In the edge feature using license plate carries out License Plate to input picture, the specific steps are as follows:
The processed filter interference noise of Gaussian mode gelatinization 2-1) is carried out to the license plate image of input, gray scale then is carried out to the image
Change;
2-2) horizontal and vertical edge is extracted to the image of gray processing by Sobel operator respectively, obtains the edge image of license plate;
Binaryzation, closed operation processing successively 2-3) are carried out to step 2-2) the license plate edge image obtained, then extract all wheels
Exterior feature finally carries out repairing treatment to the chain rupture part in profile, while obtaining edge image again using Sobel operator, thus
Extract all profiles;
Minimum circumscribed rectangle 2-4) is taken to each profile, while some not according to the tilt angle of license plate and length-width ratio primary filtration
Qualified candidate license plate, is then standardized the size of remaining candidate license plate;
2-5) sentenced using trained SVM license plate discrimination model to by the filtered remaining candidate license plate of step 2-4)
It is disconnected, obtain license plate.
4. a kind of license plate locating method combined based on color with edge feature according to claim 2 or 3, feature
It is, it is described to be referred to according to some underproof candidate license plates of tilt angle and length-width ratio primary filtration of license plate, if license plate
The absolute value of tilt angle is less than 30 degree, and the length-width ratio of license plate then retains the candidate license plate, otherwise give up this between 2-4
Candidate license plate.
5. a kind of license plate locating method combined based on color with edge feature according to claim 1, feature are existed
In the step 3-3), the picture that 70% is extracted from real license plate pictures and non-license plate pictures constitutes license plate training
Collection.
6. a kind of license plate locating method combined based on color with edge feature according to claim 1, feature are existed
In the maxPlate value takes 1.
7. a kind of license plate locating method combined based on color with edge feature according to claim 2 or 3, feature
It is, described be standardized refers to, the size of pick-up board is 136*36.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610807379.6A CN106355180B (en) | 2016-09-07 | 2016-09-07 | A kind of license plate locating method combined based on color with edge feature |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610807379.6A CN106355180B (en) | 2016-09-07 | 2016-09-07 | A kind of license plate locating method combined based on color with edge feature |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106355180A CN106355180A (en) | 2017-01-25 |
CN106355180B true CN106355180B (en) | 2019-07-02 |
Family
ID=57859712
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610807379.6A Active CN106355180B (en) | 2016-09-07 | 2016-09-07 | A kind of license plate locating method combined based on color with edge feature |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106355180B (en) |
Families Citing this family (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109239900B (en) * | 2018-11-07 | 2020-10-16 | 华东师范大学 | Full-automatic rapid focusing method for large-field acquisition of microscopic digital image |
CN109359651A (en) * | 2018-11-08 | 2019-02-19 | 东风商用车有限公司 | A kind of License Plate processor and its location processing method |
CN109871849A (en) * | 2019-01-11 | 2019-06-11 | 西安艾润物联网技术服务有限责任公司 | A kind of method and system of vehicle identification |
CN110533027B (en) * | 2019-07-22 | 2022-09-02 | 杭州未名信科科技有限公司 | Text detection and identification method and system based on mobile equipment |
CN111079744B (en) * | 2019-12-06 | 2020-09-01 | 鲁东大学 | Intelligent vehicle license plate identification method and device suitable for complex illumination environment |
CN111652222A (en) * | 2020-07-13 | 2020-09-11 | 深圳市智搜信息技术有限公司 | License plate positioning method and device, computer equipment and storage medium |
CN112488031A (en) * | 2020-12-11 | 2021-03-12 | 华能华家岭风力发电有限公司 | Safety helmet detection method based on color segmentation |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103324958A (en) * | 2013-06-28 | 2013-09-25 | 浙江大学苏州工业技术研究院 | License plate locating method based on projection method and SVM under complex background |
CN105488797A (en) * | 2015-11-25 | 2016-04-13 | 安徽创世科技有限公司 | License plate location method for HSV space |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2012145909A1 (en) * | 2011-04-28 | 2012-11-01 | 中国科学院自动化研究所 | Method for detecting tampering with color digital image based on chroma of image |
-
2016
- 2016-09-07 CN CN201610807379.6A patent/CN106355180B/en active Active
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103324958A (en) * | 2013-06-28 | 2013-09-25 | 浙江大学苏州工业技术研究院 | License plate locating method based on projection method and SVM under complex background |
CN105488797A (en) * | 2015-11-25 | 2016-04-13 | 安徽创世科技有限公司 | License plate location method for HSV space |
Non-Patent Citations (4)
Title |
---|
HSV空间和形态学处理相结合的车牌定位方法;苗姣姣 等;《电视技术》;20150302;第147页第2段-第149页倒数第2段 |
Method of license plate location based on edge detection and color information;Huili Han 等;《Proceedings 2011 International Conference on Transportation, Mechanical, and Electrical Engineering (TMEE)》;20111218;全文 |
基于边缘检测和颜色纹理直方图的车牌定位方法;郭延祥 等;《计算机科学与探索》;20140422;第720页第2段-第724倒数第2段 |
基于颜色和纹理分析的车牌定位方法;郭捷 等;《中国图象图形学报》;20020531;全文 |
Also Published As
Publication number | Publication date |
---|---|
CN106355180A (en) | 2017-01-25 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106355180B (en) | A kind of license plate locating method combined based on color with edge feature | |
CN107545239B (en) | Fake plate detection method based on license plate recognition and vehicle characteristic matching | |
CN108108761B (en) | Rapid traffic signal lamp detection method based on deep feature learning | |
CN104392205B (en) | A kind of recognition methods and system of improper license plate | |
CN105046196B (en) | Front truck information of vehicles structuring output method based on concatenated convolutional neutral net | |
CN104715239B (en) | A kind of vehicle color identification method based on defogging processing and weight piecemeal | |
Diaz-Cabrera et al. | Suspended traffic lights detection and distance estimation using color features | |
Greenhalgh et al. | Real-time detection and recognition of road traffic signs | |
CN109657632B (en) | Lane line detection and identification method | |
CN102968646B (en) | A kind of detection method of license plate based on machine learning | |
CN107103317A (en) | Fuzzy license plate image recognition algorithm based on image co-registration and blind deconvolution | |
CN103824081B (en) | Method for detecting rapid robustness traffic signs on outdoor bad illumination condition | |
CN105335743A (en) | Vehicle license plate recognition method | |
CN106022232A (en) | License plate detection method based on deep learning | |
CN106841209A (en) | One kind is based on big data self study chemical fiber wire ingot intelligence outward appearance detecting system and method | |
CN102880863B (en) | Method for positioning license number and face of driver on basis of deformable part model | |
CN104732227A (en) | Rapid license-plate positioning method based on definition and luminance evaluation | |
CN101520841A (en) | Real-time and anti-interference method for positioning license plate in high-definition TV video | |
CN106529532A (en) | License plate identification system based on integral feature channels and gray projection | |
CN104021375A (en) | Model identification method based on machine learning | |
CN105741324A (en) | Moving object detection identification and tracking method on moving platform | |
CN104598885A (en) | Method for detecting and locating text sign in street view image | |
CN106203237A (en) | The recognition methods of container-trailer numbering and device | |
CN111091544A (en) | Method for detecting breakage fault of side integrated framework of railway wagon bogie | |
CN103824091A (en) | Vehicle license plate recognition method for intelligent transportation system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 | ||
CP03 | Change of name, title or address |
Address after: Room 410, Building A3, Wuhan University Science Park, No. 3, Wudayuan 4th Road, East Lake New Technology Development Zone, Wuhan City, Hubei Province 430000 Patentee after: Wuhan Lantao Technology Co.,Ltd. Address before: Room 410, Building A3, Wuhan University Science Park, No. 3, Wudayuan 4th Road, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430223 Patentee before: WUHAN ANKE WEISHI TECHNOLOGY CO.,LTD. |
|
CP03 | Change of name, title or address |