CN103390173A - Plate number character vote algorithm based on SVM (support vector machine) confidence - Google Patents
Plate number character vote algorithm based on SVM (support vector machine) confidence Download PDFInfo
- Publication number
- CN103390173A CN103390173A CN2013103180711A CN201310318071A CN103390173A CN 103390173 A CN103390173 A CN 103390173A CN 2013103180711 A CN2013103180711 A CN 2013103180711A CN 201310318071 A CN201310318071 A CN 201310318071A CN 103390173 A CN103390173 A CN 103390173A
- Authority
- CN
- China
- Prior art keywords
- confidence
- degree
- svm
- character
- ballot
- 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
Images
Landscapes
- Character Discrimination (AREA)
Abstract
The invention discloses a plate number character vote algorithm based on SVM confidence. Aiming at a plate number recognition system based on video flows, the plate number character vote algorithm based on the SVM confidence is designed. The method comprises the steps as follows: whether voting is performed is determined according to a plate number detection result; then aiming at characteristics of plate numbers in mainland China, three SVM classifiers are trained respectively, SVMs are used for performing character recognition, and confidence corresponding to each character is output; and finally, a recognition result of each frame is voted according to the confidence, characters with the maximum confidence are selected when the voting is ended, so that a correct plate number is formed. According to the method, the plate number output finally can be guaranteed to have the highest reliability; and for conditions that the light condition is poorer, the image quality is poor, plate numbers are dirty, and the like, higher recognition rates are also realized after voting processing.
Description
Technical field
The present invention relates to computer vision technique, particularly relate to a kind of Voting Algorithm of characters on license plate based on the SVM degree of confidence
Background technology
Development along with computer technology and the information processing technology, the information processing capability of computing machine improves constantly, and computer vision technique worldwide is widely used at intelligent transportation and the electronic police system based on multimedia and pattern-recognition and artificial intelligence technology.In these application, have 96% automated system to use automatic Recognition of License Plate, the system more than 75% is to be identified as the application of core with car plate.
The precision of Vehicle License Plate Recognition System is except outside the Pass the algorithm with selecting has, and is also relevant with quality, shooting angle and the illumination condition of image.Especially for general charge station, picture quality not high (being generally D1 resolution), and camera and car plate angled, the precision of car plate identification is affected.This has two kinds of methods to solve, the one, improve picture quality, improve on-the-spot illumination condition (as adding light compensating lamp etc.), this can raise the cost undoubtedly, another kind method is, takes full advantage of the slow characteristics of charge station's speed of a motor vehicle, and same car can stop the long period, if the recognition result to multiframe is voted effectively, accuracy of identification and confidence level will improve greatly so.
By consulting existing patent, rare the car plate voting mechanism to be studied, XOR adopts and simply by occurrence number, votes, and, if the number of times that the character of identification error occurs is more, can produce wrong voting results.In order to overcome above shortcoming, existing characters on license plate Voting Algorithm based on the SVM degree of confidence, take part in a vote for the character refusal that degree of confidence is lower, and votes is directly used the degree of confidence addition, even the number of characters of correct identification is less, but, because its degree of confidence is enough high, can guarantee the accuracy of final vote result.
Summary of the invention
The invention provides a kind of Voting Algorithm of characters on license plate based on the SVM degree of confidence, the method is particularly suitable for the Vehicle License Plate Recognition System under the low speed scenes such as charge station's bayonet socket, parking lot.
To achieve these goals, the present invention includes following technical characterictic: the result that at first according to car plate, detects has judged whether that vehicle enters, and, if there is vehicle to enter, opens ballot, leave scene if the ballot number of times surpasses the threshold value of certain setting or vehicle detected, close ballot; Then utilize SVM to carry out Recognition of License Plate Characters, and export degree of confidence corresponding to each character, the characteristics that comprise Chinese character, letter and number for the China's Mainland car plate, train respectively three svm classifier devices, i.e. Chinese character sorter, alphabetic sorter and digital alphabet sorter; Finally the recognition result of every frame is voted by degree of confidence, the refusal degree of confidence participates in ballot lower than the character of the threshold value of certain setting, and while choosing poll closing, the character of degree of confidence maximum forms a correct car plate.
The described result that detects according to car plate has judged whether that vehicle enters, and, if there is vehicle to enter, opens ballot, if the ballot number of times surpasses the threshold value of certain setting or vehicle detected, leaves scene, closes ballot.
The described SVM of utilization carries out Recognition of License Plate Characters, and export degree of confidence corresponding to each character, the characteristics that comprise Chinese character, letter and number for the China's Mainland car plate, train respectively three svm classifier devices, i.e. Chinese character sorter, alphabetic sorter and digital alphabet sorter;
Described recognition result to every frame is voted by degree of confidence, and the refusal degree of confidence participates in ballot lower than the character of the threshold value of certain setting, and while choosing poll closing, the character of degree of confidence maximum forms a correct car plate.
Compared with the existing methods, the present invention is directed to the characteristics of China's Mainland car plate, train respectively three sorters, reduce target classification number, improved accuracy of identification, adopt the SVM degree of confidence to vote, and refuse the lower character of degree of confidence and take part in a vote, votes is directly used the degree of confidence addition, even the number of characters of correct identification is less, but, because its degree of confidence is enough high, can guarantee the accuracy of final vote result.
Description of drawings
Fig. 1 is the specific implementation that algorithm of the present invention is implemented use-case;
Fig. 2 is the overall flow figure of algorithm of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is only the present invention's part embodiment, rather than whole embodiment.Based on the embodiment in the present invention, those of ordinary skills, not making under the creative work prerequisite the every other embodiment that obtains, belong to the scope of protection of the invention.
Embodiment is:
(1) utilize the adboost algorithm based on the haar feature to carry out the car plate detection, if car plate detected, open ballot, and make no_plate_cnt=0, if car plate do not detected, no_plate_cnt++, no_plate_cnt reach certain threshold value or the ballot number of times reaches certain threshold value, think that vehicle leaves, stop ballot.
(2) utilize SVM to carry out character recognition, and the output degree of confidence., because the China's Mainland car plate comprises Chinese character, letter, numeral,, therefore in order to improve accuracy of identification, train respectively three sorters, i.e. Chinese character sorter, alphabetic sorter, alphanumeric sort device.Multicategory classification, the mode that adopts a plurality of two classification to vote,, according to the support vector of two classes, calculate the distance of current proper vector and support vector, and distance has reflected the degree of confidence of character recognition, and if distance is little, degree of confidence is high, and vice versa., for degree of confidence being mapped in [0,1] scope, adopt lower array function to carry out conversion.
(3) recognition result of every each character of frame checked, for the refusal of the character less than threshold value T, take part in a vote, and set maximum ballot number of times, surpass maximum ballot number of times, stop ballot.Record the voting results that each character occurs, and with the degree of confidence addition, and average as the degree of confidence of this character.Owing to the recognition result of every frame need to being added up, and do not know always to vote number of times, so need to adopt the consecutive mean method, calculate by following formula, wherein avgConfidence is average degree of confidence, and c is the degree of confidence of current recognition result, and voteNum is current ballot number of times.
(4) select each characters on license plate position, the character of degree of confidence maximum, as final voting results, and form a correct car plate, as final result output.
Easily understand, the foregoing is only preferred embodiment of the present invention, not be used for limiting spirit of the present invention and protection domain, the equivalent variations that any those of ordinary skill in the art make or replacement, within all should being considered as being encompassed in protection scope of the present invention.
Claims (4)
1. Voting Algorithm of the characters on license plate based on the SVM degree of confidence, it is characterized in that: the result that at first according to car plate, detects has judged whether that vehicle enters, if there is vehicle to enter, open ballot, leave scene if the ballot number of times surpasses the threshold value of certain setting or vehicle detected, close ballot; Then utilize SVM to carry out Recognition of License Plate Characters, and export degree of confidence corresponding to each character; Finally the recognition result of every frame is voted by degree of confidence, the refusal degree of confidence participates in ballot lower than the character of the threshold value of certain setting, and while choosing poll closing, the character of degree of confidence maximum forms a correct car plate.
2. a kind of Voting Algorithm of characters on license plate based on the SVM degree of confidence according to claim 1, it is characterized in that: the result that detects according to car plate has judged whether that vehicle enters, if there is vehicle to enter, open ballot, leave scene if the ballot number of times surpasses the threshold value of certain setting or vehicle detected, close ballot.
3. a kind of Voting Algorithm of characters on license plate based on the SVM degree of confidence according to claim 1, it is characterized in that: utilize SVM to carry out Recognition of License Plate Characters, and export degree of confidence corresponding to each character, the characteristics that comprise Chinese character, letter and number for the China's Mainland car plate, train respectively three svm classifier devices, i.e. Chinese character sorter, alphabetic sorter and digital alphabet sorter.
4. a kind of Voting Algorithm of characters on license plate based on the SVM degree of confidence according to claim 1, it is characterized in that: the recognition result to every frame is voted by degree of confidence, the refusal degree of confidence participates in ballot lower than the character of the threshold value of certain setting, and while choosing poll closing, the character of degree of confidence maximum forms a correct car plate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013103180711A CN103390173A (en) | 2013-07-24 | 2013-07-24 | Plate number character vote algorithm based on SVM (support vector machine) confidence |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2013103180711A CN103390173A (en) | 2013-07-24 | 2013-07-24 | Plate number character vote algorithm based on SVM (support vector machine) confidence |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103390173A true CN103390173A (en) | 2013-11-13 |
Family
ID=49534440
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2013103180711A Pending CN103390173A (en) | 2013-07-24 | 2013-07-24 | Plate number character vote algorithm based on SVM (support vector machine) confidence |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103390173A (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103714316A (en) * | 2013-12-10 | 2014-04-09 | 小米科技有限责任公司 | Image identification method, device and electronic equipment |
CN103824066A (en) * | 2014-03-18 | 2014-05-28 | 厦门翼歌软件科技有限公司 | Video stream-based license plate recognition method |
CN104464302A (en) * | 2014-12-22 | 2015-03-25 | 南京中兴力维软件有限公司 | License plate recognition and intelligent error correction method and system |
CN104915644A (en) * | 2015-05-29 | 2015-09-16 | 安徽四创电子股份有限公司 | License plate intelligent discrimination method based on difference data reconstruction |
CN107403130A (en) * | 2017-04-19 | 2017-11-28 | 北京粉笔未来科技有限公司 | A kind of character identifying method and character recognition device |
CN108875746A (en) * | 2018-05-17 | 2018-11-23 | 北京旷视科技有限公司 | A kind of licence plate recognition method, device, system and storage medium |
CN109697426A (en) * | 2018-12-24 | 2019-04-30 | 北京天睿空间科技股份有限公司 | Flight based on multi-detector fusion shuts down berth detection method |
CN110084232A (en) * | 2018-01-25 | 2019-08-02 | 浙江宇视科技有限公司 | The recognition methods of chinese character, device and terminal device in license plate |
CN111461111A (en) * | 2020-03-03 | 2020-07-28 | 华南理工大学 | Multiframe license plate recognition optimization method based on random forest |
CN112907832A (en) * | 2021-01-19 | 2021-06-04 | 浙江大华技术股份有限公司 | Processing method of refueling event, video processing device and storage medium |
CN113723422A (en) * | 2021-09-08 | 2021-11-30 | 重庆紫光华山智安科技有限公司 | License plate information determination method, system, device and medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201134166Y (en) * | 2007-10-24 | 2008-10-15 | 北京万集科技有限责任公司 | Number plate of vehicle recognition system |
CN101630367A (en) * | 2009-07-31 | 2010-01-20 | 北京科技大学 | Rejection method for identifying handwritten character based on multiple classifiers |
CN102346847A (en) * | 2011-09-26 | 2012-02-08 | 青岛海信网络科技股份有限公司 | License plate character recognizing method of support vector machine |
WO2012074526A1 (en) * | 2010-12-02 | 2012-06-07 | 3M Innovative Properties Company | Methods and systems for enhancing read accuracy in automated license plate reader systems |
CN102880859A (en) * | 2012-08-30 | 2013-01-16 | 华南理工大学 | Method for recognizing number plate |
-
2013
- 2013-07-24 CN CN2013103180711A patent/CN103390173A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN201134166Y (en) * | 2007-10-24 | 2008-10-15 | 北京万集科技有限责任公司 | Number plate of vehicle recognition system |
CN101630367A (en) * | 2009-07-31 | 2010-01-20 | 北京科技大学 | Rejection method for identifying handwritten character based on multiple classifiers |
WO2012074526A1 (en) * | 2010-12-02 | 2012-06-07 | 3M Innovative Properties Company | Methods and systems for enhancing read accuracy in automated license plate reader systems |
CN102346847A (en) * | 2011-09-26 | 2012-02-08 | 青岛海信网络科技股份有限公司 | License plate character recognizing method of support vector machine |
CN102880859A (en) * | 2012-08-30 | 2013-01-16 | 华南理工大学 | Method for recognizing number plate |
Non-Patent Citations (2)
Title |
---|
张伟伟 等: "基于DPM的自然场景下汉字识别方法", 《计算机应用研究》 * |
黎涛: "车牌字符识别技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103714316B (en) * | 2013-12-10 | 2017-03-01 | 小米科技有限责任公司 | Image-recognizing method, device and electronic equipment |
CN103714316A (en) * | 2013-12-10 | 2014-04-09 | 小米科技有限责任公司 | Image identification method, device and electronic equipment |
CN103824066A (en) * | 2014-03-18 | 2014-05-28 | 厦门翼歌软件科技有限公司 | Video stream-based license plate recognition method |
CN103824066B (en) * | 2014-03-18 | 2017-10-10 | 厦门翼歌软件科技有限公司 | A kind of licence plate recognition method based on video flowing |
CN104464302A (en) * | 2014-12-22 | 2015-03-25 | 南京中兴力维软件有限公司 | License plate recognition and intelligent error correction method and system |
CN104915644A (en) * | 2015-05-29 | 2015-09-16 | 安徽四创电子股份有限公司 | License plate intelligent discrimination method based on difference data reconstruction |
CN104915644B (en) * | 2015-05-29 | 2018-01-19 | 安徽四创电子股份有限公司 | A kind of car plate intelligent distinguishing method rebuild based on difference data |
CN107403130A (en) * | 2017-04-19 | 2017-11-28 | 北京粉笔未来科技有限公司 | A kind of character identifying method and character recognition device |
CN110084232A (en) * | 2018-01-25 | 2019-08-02 | 浙江宇视科技有限公司 | The recognition methods of chinese character, device and terminal device in license plate |
CN110084232B (en) * | 2018-01-25 | 2021-01-29 | 浙江宇视科技有限公司 | Method and device for recognizing Chinese characters in license plate and terminal equipment |
CN108875746A (en) * | 2018-05-17 | 2018-11-23 | 北京旷视科技有限公司 | A kind of licence plate recognition method, device, system and storage medium |
CN108875746B (en) * | 2018-05-17 | 2023-02-17 | 北京旷视科技有限公司 | License plate recognition method, device and system and storage medium |
CN109697426B (en) * | 2018-12-24 | 2019-10-18 | 北京天睿空间科技股份有限公司 | Flight based on multi-detector fusion shuts down berth detection method |
CN109697426A (en) * | 2018-12-24 | 2019-04-30 | 北京天睿空间科技股份有限公司 | Flight based on multi-detector fusion shuts down berth detection method |
CN111461111A (en) * | 2020-03-03 | 2020-07-28 | 华南理工大学 | Multiframe license plate recognition optimization method based on random forest |
CN111461111B (en) * | 2020-03-03 | 2024-01-05 | 华南理工大学 | Multi-frame license plate recognition optimization method based on random forest |
CN112907832A (en) * | 2021-01-19 | 2021-06-04 | 浙江大华技术股份有限公司 | Processing method of refueling event, video processing device and storage medium |
CN113723422A (en) * | 2021-09-08 | 2021-11-30 | 重庆紫光华山智安科技有限公司 | License plate information determination method, system, device and medium |
CN113723422B (en) * | 2021-09-08 | 2023-10-17 | 重庆紫光华山智安科技有限公司 | License plate information determining method, system, equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103390173A (en) | Plate number character vote algorithm based on SVM (support vector machine) confidence | |
CN102346847B (en) | License plate character recognizing method of support vector machine | |
CN102722704A (en) | Method and system for recognizing vehicle license plate by integrating video dynamic tracking | |
CN103971097A (en) | Vehicle license plate recognition method and system based on multiscale stroke models | |
CN104766042A (en) | Method and apparatus for and recognizing traffic sign board | |
CN111161543A (en) | Automatic snapshot method and system for bus front violation behavior based on image recognition | |
CN107016362B (en) | Vehicle weight recognition method and system based on vehicle front windshield pasted mark | |
CN112966631A (en) | License plate detection and identification system and method under unlimited security scene | |
CN101937508A (en) | License plate localization and identification method based on high-definition image | |
EA016732B1 (en) | System and method for detection of average speed of vehicles for traffic control | |
CN103279756A (en) | Vehicle detecting analysis system and detecting analysis method thereof based on integrated classifier | |
CN104408475B (en) | A kind of licence plate recognition method and car license recognition equipment | |
CN105512660A (en) | License number identification method and device | |
CN102637257A (en) | Video-based detection and recognition system and method of vehicles | |
CN104537360A (en) | Method and system for detecting vehicle violation of not giving way | |
CN102880859A (en) | Method for recognizing number plate | |
CN104992173A (en) | Symbol recognition method and system used for medical report | |
CN103246893A (en) | ECOC (European Conference on Optical Communication) encoding classification method based on rejected random subspace | |
CN105354573A (en) | Container license plate identification method and system | |
CN110009058A (en) | A kind of parking lot Vehicle License Plate Recognition System and method | |
CN103971092A (en) | Facial trajectory tracking method | |
KR101794632B1 (en) | The logo pattern recognizing method in number plate of electric vehicle | |
CN112085018A (en) | License plate recognition system based on neural network | |
CN103106400B (en) | A kind of method for detecting human face and device | |
Xu et al. | A machine learning approach to vehicle occupancy detection |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20131113 |