CN104915644B - A kind of car plate intelligent distinguishing method rebuild based on difference data - Google Patents
A kind of car plate intelligent distinguishing method rebuild based on difference data Download PDFInfo
- Publication number
- CN104915644B CN104915644B CN201510286086.3A CN201510286086A CN104915644B CN 104915644 B CN104915644 B CN 104915644B CN 201510286086 A CN201510286086 A CN 201510286086A CN 104915644 B CN104915644 B CN 104915644B
- Authority
- CN
- China
- Prior art keywords
- plate
- character
- car
- vehicle
- plate number
- 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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Discrimination (AREA)
- Traffic Control Systems (AREA)
- Vehicle Waterproofing, Decoration, And Sanitation Devices (AREA)
Abstract
The present invention relates to the car plate intelligent distinguishing method rebuild based on difference data, including vehicle is taken pictures；Car plate photo is inputted in Vehicle License Plate Recognition System and is identified, determines that six characters that the number-plate number is included are the letter or number corresponding in confidence level maximum；Six characters are arranged by the order of confidence value from big to small；Number-plate number confidence threshold value is set, the character more than the number-plate number of the threshold value is credible character, and the character less than the number-plate number of the threshold value is suspicious character；Car record is crossed with reference in bayonet system database, determines that car plate integrates the confidence level highest number-plate number using difference data method for reconstructing；Number-plate number combination information of vehicles is crossed into car record as one to be stored in the database of bayonet system.The present invention can effectively improve the adaptivity and robustness of Vehicle License Plate Recognition System, greatly mitigate the pressure of vehicle license location technique, Character segmentation technology and character recognition technologies, improve the accuracy of number-plate number identification.
Description
Technical field
The present invention relates to intelligent transportation and car plate intelligent identification technology field, it is especially a kind of based on difference data rebuild
Car plate intelligent distinguishing method.
Background technology
In recent years, as expanding economy, the material life of people need increasingly to improve, private motor vehicles owning amount also exists
Increase in geometry quantity, the popularization of vehicle turns into inevitable trend.But thing followed traffic congestion, increasing environmental pollution,
The problems such as traffic accident takes place frequently also becomes increasingly conspicuous, and the increase of automobile quantity has become an important factor for restricting urban development, because
This needs the road network of Erecting and improving to alleviate the demand of road traffic growth, only greatly develops intelligent transportation system,
It is possible to the traffic problems for really solving getting worse.How accurately to carry out Intelligent Recognition to license plate is vehicle detecting system
In an important step, it traffic monitoring and control in occupy an important position, be to realize traffic administration modernization and intelligence
The important prerequisite of energyization, the identification problem of license plate turn into the emphasis studied in Modern Traffic engineering field and focus it
One.
Due to being influenceed by unfixed etc. the factor in environment, the vehicle of vehicle to be identified, car plate position, to car plate in picture
Positioning belt carry out certain difficulty, the car plates of some vehicles is infected with by dust, soil so that car plate is contaminated in itself.In addition, also have
Some vehicles due to running time it is longer so that license plate number segmental defect, these situations can all influence system to the number-plate number
Accuracy is identified, results even in misrecognition.It can be seen that to improve the accuracy of Car license recognition, simple relies on License Plate skill
Art, Character segmentation technology and image recognition technology are extremely difficult to.Therefore, there is an urgent need to design one kind to have good robust
Property and adaptivity Intelligent license-plate of vehicle method of discrimination, realize to have certain noise, deformation in addition missing characters on license plate remain to just
Really identification, so as to mitigate the pressure of vehicle license location technique and Character segmentation technology.
The content of the invention
It is an object of the invention to provide a kind of identification that can improve the car plate for having pollution in itself or the car plate of defect is accurate
True property, so as to improve the anti-interference of Vehicle License Plate Recognition System, mitigate the base of the pressure of vehicle license location technique and Character segmentation technology
In the car plate intelligent distinguishing method that difference data is rebuild.
To achieve the above object, present invention employs following technical scheme：A kind of car plate intelligence rebuild based on difference data
Can method of discrimination, the step of this method includes following order：
（1）Vehicle is taken pictures by the camera in the bayonet system of front end；
（2）Photo to be identified containing vehicle license plate number is inputted into the car plate based on computer image processing technology to know
It is identified in other system, determines that six characters that the number-plate number is included are six letters corresponding in confidence level maximum
Or numeral；
（3）Six characters of the number-plate number identified are arranged by the order of confidence value from big to small；
（4）Number-plate number confidence threshold value is set, the character more than the number-plate number of the threshold value is credible character, less than this
The character of the number-plate number of threshold value is suspicious character；
（5）Car record is crossed with reference in bayonet system database, determines that car plate synthesis is put using difference data method for reconstructing
The reliability highest number-plate number；
（6）The number-plate number of determination is combined into car plate color, the vehicle face by being acquired to vehicle image feature extraction
Color, vehicle logo information are crossed car record as one and are stored in the database of bayonet system.
Camera in the front end bayonet system to vehicle take pictures the vehicle referred to Jing Guo bayonet system, from car
Front end or rear end taken pictures；For virtual bayonet socket, by the vehicle photo upload of shooting to background data center, in rear number of units
Six characters that the number-plate number is carried out according to center extract operation；For entity bayonet socket, by the vehicle photo of shooting in bayonet system
The operation of characters on license plate extraction is carried out, obtained differentiation result is sent to background data center and stored.
When number-plate number information is identified the Vehicle License Plate Recognition System, vehicle photo progress image is located in advance first
Reason, License Plate and Character segmentation, Recognition of License Plate Characters operation is then carried out, so that it is determined that the character number of each number-plate number exists
Corresponding letter or number information during confidence level maximum.
When the confidence threshold value to the number-plate number is set, first using the method for mathematical statistics, license plate number is determined
Confidence value when code identification accuracy is high, then confidence value in number-plate number character identification result is more than to the car of the threshold value
For trade mark code character as credible character, confidence value is less than the number-plate number character of the threshold value as suspicious character.
The difference data method for reconstructing comprises the following steps：
（1）Credible character in the number-plate number to be identified is extracted, makes the position of suspicious character in the number-plate number
Vacancy forms space character, the number-plate number is changed into the characters on license plate string containing space character；Then retrieved in bayonet socket database
A period of time interior energy is with being somebody's turn to do the license plate number of the characters on license plate string fuzzy matching containing space character and its being shot from car plate to be identified
The car of crossing of time recently is recorded, and license plate number different with to be identified vehicle vehicle, color in the vehicle retrieved is deleted；
（2）According to the license plate number retrieved and its car record is crossed, and the time difference that the car plate to be differentiated occurs, and
The distance between two shooting places, its average speed is calculated, if the average speed is more than what common vehicle travelled in two places
Average speed higher limit, then the car plate retrieved is deleted, otherwise retain the car plate；
（3）Confidence level highest character is added for retrieval in the suspicious character that the number-plate number to be identified is included
In characters on license plate string, average speed, and and common vehicle are obtained with car information of crossing to be discriminated according to the car record of crossing retrieved
In two places, the average speed upper limit of traveling compares, and car plate is screened；
（4）Judge whether remaining suspicious character is empty, if suspicious character is sky, differentiation terminates, and determines that car plate integrates
The confidence level highest number-plate number；Otherwise, confidence level highest character adds the car for retrieval in the remaining suspicious character of taking-up
In board character string, repeat step（3）.
For the span of the confidence level to be less than 1 constant more than zero, the confidence threshold value is 80%.
As shown from the above technical solution, the advantage of the invention is that：First, it will be identified first based on image processing techniques
The number-plate number is arranged by the order of confidence level from big to small, and sets number-plate number confidence threshold value, then in conjunction with bayonet socket
Car record is crossed in system database, the method rebuild using difference data determines that car plate integrates confidence level highest license plate number
Code, and this analysis result is stored in bayonet system, it is possible to increase itself there is the identification of the car plate of pollution or the car plate of defect accurate
True property, so as to improve the anti-interference of Vehicle License Plate Recognition System；Second, it is of the invention to be combined with traditional image processing techniques, can
Effectively improve the adaptivity and robustness of Vehicle License Plate Recognition System, greatly mitigate vehicle license location technique, Character segmentation technology and
The pressure of character recognition technologies, further improve the accuracy of number-plate number identification.
Brief description of the drawings
Fig. 1 is flow chart of the method for the present invention；
Fig. 2 is the flow chart of difference data method for reconstructing in the present invention.
Embodiment
As shown in figure 1, a kind of car plate intelligent distinguishing method rebuild based on difference data, is comprised the following steps：（1）Pass through
Camera in the bayonet system of front end is taken pictures to vehicle；（2）Photo to be identified containing vehicle license plate number is inputted
It is identified in Vehicle License Plate Recognition System based on computer image processing technology, determines that six characters that the number-plate number is included are
Six corresponding letter or numbers in confidence level maximum；（3）By six characters of the number-plate number identified by confidence level
The order of value from big to small is arranged；（4）Number-plate number confidence threshold value is set, more than the character of the number-plate number of the threshold value
For credible character, the character less than the number-plate number of the threshold value is suspicious character；（5）Car is crossed with reference in bayonet system database
Record, determine that car plate integrates the confidence level highest number-plate number using difference data method for reconstructing；（6）By the number-plate number of determination
It is used as one with reference to the vehicle color by being acquired to vehicle image feature extraction, car plate color, vehicle logo information and crosses car
Record is stored in the database of bayonet system.
As shown in figure 1, the camera in the front end bayonet system take pictures referring to by bayonet system to vehicle
Vehicle, taken pictures from the front end of car or rear end；For virtual bayonet socket, by the vehicle photo upload of shooting into back-end data
The heart, six characters that the number-plate number is carried out in background data center extract operation；For entity bayonet socket, by the vehicle photo of shooting
The operation of characters on license plate extraction is carried out in bayonet system, obtained differentiation result is sent to background data center and stored.
As shown in figure 1, when number-plate number information is identified the Vehicle License Plate Recognition System, vehicle photo is carried out first
Image preprocessing, License Plate and Character segmentation, Recognition of License Plate Characters operation is then carried out, so that it is determined that each number-plate number
Character number corresponding letter or number information in confidence level maximum.
As shown in figure 1, when the confidence threshold value to the number-plate number is set, first using the method for mathematical statistics,
Confidence value when number-plate number identification accuracy is high is determined, is then more than confidence value in number-plate number character identification result
For the number-plate number character of the threshold value as credible character, confidence value is less than the number-plate number character of the threshold value as suspicious word
Symbol.The span of the confidence level is is less than 1 constant more than zero, and herein, the confidence threshold value is 80%.
As shown in Fig. 2 the difference data method for reconstructing comprises the following steps：
（1）Credible character in the number-plate number to be identified is extracted, makes the position of suspicious character in the number-plate number
Vacancy forms space character, the number-plate number is changed into the characters on license plate string containing space character；Then retrieved in bayonet socket database
A period of time interior energy is with being somebody's turn to do the license plate number of the characters on license plate string fuzzy matching containing space character and its being shot from car plate to be identified
The car of crossing of time recently is recorded, and license plate number different with to be identified vehicle vehicle, color in the vehicle retrieved is deleted；
（2）According to the license plate number retrieved and its car record is crossed, and the time difference that the car plate to be differentiated occurs, and
The distance between two shooting places, its average speed is calculated, if the average speed is more than what common vehicle travelled in two places
Average speed higher limit, then the car plate retrieved is deleted, otherwise retain the car plate；For example in city, average speed is set
It is 150 km/h to spend higher limit；
（3）Confidence level highest character is added for retrieval in the suspicious character that the number-plate number to be identified is included
In characters on license plate string, average speed, and and common vehicle are obtained with car information of crossing to be discriminated according to the car record of crossing retrieved
In two places, the average speed upper limit of traveling compares, and car plate is screened；
（4）Judge whether remaining suspicious character is empty, if suspicious character is sky, differentiation terminates, and determines that car plate integrates
The confidence level highest number-plate number；Otherwise, confidence level highest character adds the car for retrieval in the remaining suspicious character of taking-up
In board character string, repeat step（3）.
In summary, the present invention can correctly be identified to the characters on license plate image for having certain noise or deformation, so as to
The accuracy of Car license recognition is effectively improved, the adaptivity and robustness of Vehicle License Plate Recognition System can be effectively improved, is greatly mitigated
The pressure of vehicle license location technique, Character segmentation technology and character recognition technologies.
Claims (5)
1. a kind of car plate intelligent distinguishing method rebuild based on difference data, the step of this method includes following order：
（1）Vehicle is taken pictures by the camera in the bayonet system of front end；
（2）Photo to be identified containing vehicle license plate number is inputted into the Car license recognition system based on computer image processing technology
It is identified in system, determines that six characters that the number-plate number is included are six letters or number corresponding in confidence level maximum
Word；
（3）Six characters of the number-plate number identified are arranged by the order of confidence value from big to small；
（4）Number-plate number confidence threshold value is set, the character more than the number-plate number of the threshold value is credible character, less than the threshold value
The character of the number-plate number be suspicious character；
（5）Car record is crossed with reference in bayonet system database, determines that car plate integrates confidence level using difference data method for reconstructing
The highest number-plate number；
（6）The number-plate number of determination is combined into the car plate color, vehicle color, car by being acquired to vehicle image feature extraction
Logo information is crossed car record as one and is stored in the database of bayonet system；
The difference data method for reconstructing comprises the following steps：
（1）Credible character in the number-plate number to be identified is extracted, makes the position vacancy of suspicious character in the number-plate number
Space character is formed, the number-plate number is changed into the characters on license plate string containing space character；Then one section is retrieved in bayonet socket database
Time interior energy is with being somebody's turn to do the license plate number of the characters on license plate string fuzzy matching containing space character and its from car plate shooting time to be identified
Nearest crosses car record, and license plate number different with to be identified vehicle vehicle, color in the vehicle retrieved is deleted；
（2）According to the license plate number retrieved and its car record is crossed, and the time difference that the car plate to be differentiated occurs, and two bats
According to the distance between place, its average speed is calculated, if the average speed is averaged more than common vehicle what two places travelled
Speed limit value, then the car plate retrieved is deleted, otherwise retain the car plate；
（3）Confidence level highest character adds the car plate for retrieval in the suspicious character that the number-plate number to be identified is included
In character string, average speed is obtained with car information of crossing to be discriminated according to the car record of crossing retrieved, and with common vehicle two
The average speed upper limit of ground traveling compares, and car plate is screened；
（4）Judge whether remaining suspicious character is empty, if suspicious character is sky, differentiation terminates, and determines that car plate integrates confidence
Spend the highest number-plate number；Otherwise, confidence level highest character adds the car plate word for retrieval in the remaining suspicious character of taking-up
In symbol string, repeat step（3）.
2. the car plate intelligent distinguishing method according to claim 1 rebuild based on difference data, it is characterised in that：Before described
Camera in the bayonet system of end to vehicle take pictures the vehicle referred to Jing Guo bayonet system, enters from the front end of car or rear end
Row is taken pictures；For virtual bayonet socket, by the vehicle photo upload of shooting to background data center, car plate is carried out in background data center
Six characters extraction operation of number；For entity bayonet socket, the vehicle photo of shooting is subjected to characters on license plate in bayonet system and carried
The operation taken, obtained differentiation result is sent to background data center and stored.
3. the car plate intelligent distinguishing method according to claim 1 rebuild based on difference data, it is characterised in that：The car
When number-plate number information is identified board identifying system, vehicle photo is subjected to image preprocessing, License Plate and word first
Symbol segmentation, then carries out Recognition of License Plate Characters operation, so that it is determined that the character number of each number-plate number is in confidence level maximum pair
The letter or number information answered.
4. the car plate intelligent distinguishing method according to claim 1 rebuild based on difference data, it is characterised in that：To car
When the confidence threshold value of trade mark code is set, first using the method for mathematical statistics, determine that number-plate number identification accuracy is high
When confidence value, then using confidence value in number-plate number character identification result be more than the threshold value number-plate number character as
Credible character, confidence value are less than the number-plate number character of the threshold value as suspicious character.
5. the car plate intelligent distinguishing method according to claim 4 rebuild based on difference data, it is characterised in that：It is described to put
The span of reliability is the constant less than 1, confidence threshold value 80% more than zero.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510286086.3A CN104915644B (en) | 2015-05-29 | 2015-05-29 | A kind of car plate intelligent distinguishing method rebuild based on difference data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510286086.3A CN104915644B (en) | 2015-05-29 | 2015-05-29 | A kind of car plate intelligent distinguishing method rebuild based on difference data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104915644A CN104915644A (en) | 2015-09-16 |
| CN104915644B true CN104915644B (en) | 2018-01-19 |
Family
ID=54084696
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510286086.3A Active CN104915644B (en) | 2015-05-29 | 2015-05-29 | A kind of car plate intelligent distinguishing method rebuild based on difference data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104915644B (en) |
Families Citing this family (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105139659A (en) * | 2015-10-09 | 2015-12-09 | 东方网力科技股份有限公司 | Vehicle license plate recognition method and device |
| CN105321352A (en) * | 2015-11-16 | 2016-02-10 | 深圳市哈工大交通电子技术有限公司 | A motor vehicle license plate blocking violation detection and evidence obtaining method |
| CN106780886B (en) * | 2016-12-16 | 2020-02-04 | 深圳市捷顺科技实业股份有限公司 | Vehicle identification system and vehicle entrance and exit identification method |
| CN106652469B (en) * | 2016-12-16 | 2019-07-30 | 深圳市捷顺科技实业股份有限公司 | A kind of vehicle marches into the arena, vehicle appearance licence plate recognition method and Vehicle License Plate Recognition System |
| CN108664957B (en) | 2017-03-31 | 2021-08-24 | 杭州海康威视数字技术股份有限公司 | License plate number matching method and device, and character information matching method and device |
| CN107273848A (en) * | 2017-06-13 | 2017-10-20 | 成都尽知致远科技有限公司 | Video analysis method based on Number Reorganization |
| CN109063702B (en) * | 2018-08-08 | 2022-05-17 | 北京旷视科技有限公司 | License plate recognition method, device, equipment and storage medium |
| CN110491133B (en) * | 2019-08-08 | 2020-10-16 | 善泊科技(珠海)有限公司 | Vehicle information correction system and method based on confidence |
| CN111652234A (en) * | 2020-06-04 | 2020-09-11 | 上海眼控科技股份有限公司 | License plate recognition method and system, computer equipment and readable storage medium |
| CN113593254B (en) * | 2021-07-30 | 2023-02-03 | 青岛海信智慧生活科技股份有限公司 | Parking space management method, equipment and system |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102262726A (en) * | 2011-06-17 | 2011-11-30 | 西安电子科技大学 | Multi-core license plate recognition system based on field programmable gate array (FPGA) |
| CN103390173A (en) * | 2013-07-24 | 2013-11-13 | 佳都新太科技股份有限公司 | Plate number character vote algorithm based on SVM (support vector machine) confidence |
| CN103927880A (en) * | 2014-04-28 | 2014-07-16 | 浙江宇视科技有限公司 | Vehicle license plate recognizing and matching method and device |
-
2015
- 2015-05-29 CN CN201510286086.3A patent/CN104915644B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102262726A (en) * | 2011-06-17 | 2011-11-30 | 西安电子科技大学 | Multi-core license plate recognition system based on field programmable gate array (FPGA) |
| CN103390173A (en) * | 2013-07-24 | 2013-11-13 | 佳都新太科技股份有限公司 | Plate number character vote algorithm based on SVM (support vector machine) confidence |
| CN103927880A (en) * | 2014-04-28 | 2014-07-16 | 浙江宇视科技有限公司 | Vehicle license plate recognizing and matching method and device |
Non-Patent Citations (1)
| Title |
|---|
| 质量退化的车牌字符分割方法;李文举等;《计算机辅助设计与图形学学报》;20040531;第16卷(第5期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104915644A (en) | 2015-09-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104915644B (en) | A kind of car plate intelligent distinguishing method rebuild based on difference data | |
| CN107729818B (en) | Multi-feature fusion vehicle re-identification method based on deep learning | |
| CN109637151B (en) | Method for identifying illegal driving of emergency lane on highway | |
| CN103136528B (en) | A kind of licence plate recognition method based on dual edge detection | |
| CN104966071B (en) | It is a kind of based on the night car plate detection of infrared light filling and recognition methods and device | |
| CN105702048B (en) | Highway front truck illegal road occupation identifying system based on automobile data recorder and method | |
| CN110619279B (en) | Road traffic sign instance segmentation method based on tracking | |
| CN110879950A (en) | Multi-stage target classification and traffic sign detection method and device, equipment and medium | |
| US20160117562A1 (en) | Traffic sign recognizing apparatus and operating method thereof | |
| CN103116748B (en) | Based on the method and system of road surface identification identification violation driving behavior | |
| CN101937508A (en) | License plate localization and identification method based on high-definition image | |
| CN107506368A (en) | The determination method and device of one species case suspected vehicles | |
| CN106940784A (en) | A kind of bus detection and recognition methods and system based on video | |
| CN105894819A (en) | Fake-license-plate vehicle identification method based on twice verification | |
| CN111369801B (en) | Vehicle identification method, device, equipment and storage medium | |
| CN112785850A (en) | Method and device for identifying vehicle lane change without lighting | |
| CN106529540A (en) | Gray projection-based license plate character segmentation method | |
| CN112115939A (en) | Vehicle license plate recognition method and device | |
| CN107358236A (en) | A kind of number-plate number identifying system and method based on camera device | |
| CN112115800A (en) | Vehicle combination recognition system and method based on deep learning target detection | |
| CN113408364B (en) | Temporary license plate recognition method, system, device and storage medium | |
| CN203020184U (en) | Electric vehicle lane marking line detecting device based on double charge coupled devices (CCD) | |
| CN103679209A (en) | Sparse theory based character recognition method | |
| CN106778765B (en) | License plate recognition method and device | |
| CN107622667B (en) | Method and system for detecting altered license plate number |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |