CN105631470A - Method and system for verifying license plate type - Google Patents
Method and system for verifying license plate type Download PDFInfo
- Publication number
- CN105631470A CN105631470A CN201510969964.1A CN201510969964A CN105631470A CN 105631470 A CN105631470 A CN 105631470A CN 201510969964 A CN201510969964 A CN 201510969964A CN 105631470 A CN105631470 A CN 105631470A
- Authority
- CN
- China
- Prior art keywords
- character
- type
- car plate
- plate
- model
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Traffic Control Systems (AREA)
- Character Discrimination (AREA)
Abstract
The invention discloses a method and a system for verifying a license plate type. The method comprises steps: a license plate image is acquired; the license plate image is recognized according to a license plate type recognition classifier, and a roughly-classified type of the license plate image is determined by the recognition result according to a predetermined condition; a license plate character template corresponding to the roughly-classified type is called to carry out character segmentation on the license plate in the license plate image; according to the roughly-classified type, characters at a license plate difference position are determined, and a character model corresponding to the characters at the license plate difference position is used for recognizing the characters at the license plate difference position; the recognition result is compared with a threshold of the character model corresponding to the characters at the license plate difference position to determine the type of the license plate image. Thus, the license plate type can be quickly and accurately verified, and the problem that the license plate is wrongly recognized as the type recognition in multi-type license plate recognition is inaccurate is solved to a certain degree.
Description
Technical field
The present invention relates to image processing field, particularly to the method and system of a kind of car plate type approval.
Background technology
Along with sharply increasing of China's automobile quantity, the range of application of license plate recognition technology is also more and more wider, current license plate recognition technology has been largely used to the occasions such as wagon flow monitoring, access and exit control, electronic charging, mobile check, overspeed detection, but still there are some problems, such as the car plate leakage caused due to picture quality identifies, similar character mistake; Owing to the more type identification caused of car plate type is inaccurate, polymorphic type Car license recognition rate is not high or real-time is poor etc.
Vehicle License Plate Recognition System mainly includes with lower part: License Plate, license plate character segmentation, Recognition of License Plate Characters. Wherein car plate type identification may be considered the sub-step of license plate character segmentation, first determines corresponding car plate type, then adopts corresponding characters on license plate template that car plate is carried out character cutting again. Car plate type identification is a vital link in Car license recognition, and the accuracy of car plate type identification determines the discrimination of polymorphic type car plate to a certain extent. Therefore, how to carry out car plate type approval accurately, be those skilled in the art's technical issues that need to address.
Summary of the invention
It is an object of the invention to provide the method and system of a kind of car plate type approval, it is possible to verify the type of car plate fast and accurately, to some extent solve the problem causing Car license recognition mistake in polymorphic type Car license recognition because type identification is inaccurate.
For solving above-mentioned technical problem, the present invention provides a kind of method of car plate type approval, including:
Obtain license plate image;
By car plate type identification grader, described license plate image is identified, determines the rough sort type of described license plate image according to recognition result according to predetermined condition;
Call the characters on license plate template corresponding with described rough sort type and the car plate in described license plate image is carried out character cutting;
According to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding;
The threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the type of described license plate image.
Wherein, call the characters on license plate template corresponding with described rough sort type and the car plate in described license plate image is carried out character cutting, including:
When described rough sort type is black board, then call blue-black board Character mother plate and car plate is carried out character cutting by embassy's board Character mother plate;
When described rough sort type is monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate or alert board, then call the monolayer yellow card Character mother plate of correspondence, monolayer army board Character mother plate, monolayer People's Armed Police's characters on license plate template or alert board Character mother plate and car plate is carried out character cutting;
When described rough sort type is double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card, then call the double-deck army board Character mother plate of correspondence, double-deck People's Armed Police's characters on license plate template or double-deck yellow card Character mother plate and car plate is carried out character cutting;
When described rough sort type is blue board, then calls blue-black board Character mother plate and car plate is carried out character cutting.
Wherein, described character model includes:
Utilize the two LDA model classified and the SVM models that " making " trains;
Utilize the two LDA model classified and the SVM models that " WJ " trains;
Utilize the two LDA model classified and the SVM models that " police " trains;
Utilize the 12 LDA model classified and the SVM models that " 12 army's board initials " trains.
Wherein, the threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the accurate type of described license plate image, including:
LDA model that character according to described car plate differential position is corresponding and SVM model, the LDA model confidence of output and SVM model confidence;
The LDA model threshold that described LDA model confidence is corresponding with the character of described car plate differential position is compared, if more than LDA model threshold, then obtains the first result;
The SVM model threshold that described SVM model confidence is corresponding with the character of described car plate differential position is compared, if more than SVM model threshold, then obtains the second result;
When meeting described first result and described second result simultaneously, then the type of described license plate image is the type that described LDA model is corresponding with SVM model.
Wherein, also include:
Regularly described character model is safeguarded.
The present invention provides the system of a kind of car plate type approval, including:
Acquisition module, is used for obtaining license plate image;
Rough sort module, for described license plate image being identified by car plate type identification grader, determines the rough sort type of described license plate image according to recognition result according to predetermined condition;
Cutting module, carries out character cutting for calling the characters on license plate template corresponding with described rough sort type to the car plate in described license plate image;
Character recognition module, for according to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding;
Determination type module, for comparing the threshold value of character model corresponding with the character of described car plate differential position for recognition result, it is determined that the type of described license plate image.
Wherein, described cutting module includes:
First cutting unit, for being black board when described rough sort type, then calls blue-black board Character mother plate and car plate is carried out character cutting by embassy's board Character mother plate;
Second cutting unit, for being monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate or alert board when described rough sort type, then call the monolayer yellow card Character mother plate of correspondence, monolayer army board Character mother plate, monolayer People's Armed Police's characters on license plate template or alert board Character mother plate and car plate is carried out character cutting;
3rd cutting unit, for being double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card when described rough sort type, then calls the double-deck army board Character mother plate of correspondence, double-deck People's Armed Police's characters on license plate template or double-deck yellow card Character mother plate and car plate is carried out character cutting;
4th cutting unit, for being blue board when described rough sort type, then calls blue-black board Character mother plate and car plate is carried out character cutting.
Wherein, this system includes:
First training module, is used for the two LDA model classified and the SVM models utilizing " making " to train;
Second training module, is used for the two LDA model classified and the SVM models utilizing " WJ " to train;
3rd training module, is used for the two LDA model classified and the SVM models utilizing " police " to train;
4th training module, is used for the 12 LDA model classified and the SVM models utilizing " 12 army's board initials " to train.
Wherein, described determination type module includes:
Output unit, for the LDA model corresponding according to the character of described car plate differential position and SVM model, the LDA model confidence of output and SVM model confidence;
First comparing unit, for the LDA model threshold that described LDA model confidence is corresponding with the character of described car plate differential position being compared, if more than LDA model threshold, then obtains the first result;
Second comparing unit, for the SVM model threshold that described SVM model confidence is corresponding with the character of described car plate differential position being compared, if more than SVM model threshold, then obtains the second result;
Type determining units, for when meeting described first result and described first result simultaneously, then the type of described license plate image is the type that described LDA model is corresponding with SVM model.
Wherein, also include:
Maintenance module, for regularly safeguarding described character model.
The method and system of car plate type approval provided by the present invention, including: obtain license plate image; By car plate type identification grader, described license plate image is identified, determines the rough sort type of described license plate image according to recognition result according to predetermined condition; Call the characters on license plate template corresponding with described rough sort type and the car plate in described license plate image is carried out character cutting; According to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding; The threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the type of described license plate image;
The type of car plate is classified by the method by using car plate type identification grader, utilize the characters on license plate template corresponding with classification results that the character in car plate is carried out cutting, the character of car plate differential position is identified by the character model utilizing the character of car plate differential position corresponding, and determine that whether the type that car plate type identification grader obtains is accurate according to recognition result, and finally determine the accurate type of car plate according to recognition result; Namely the method is after the type of car plate is classified by car plate type identification grader, and sorted result is verified again; Therefore, it is possible to verify the type of car plate fast and accurately, to some extent solve the problem causing Car license recognition mistake in polymorphic type Car license recognition because type identification is inaccurate.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, the accompanying drawing used required in embodiment or description of the prior art will be briefly described below, apparently, accompanying drawing in the following describes is only embodiments of the invention, for those of ordinary skill in the art, under the premise not paying creative work, it is also possible to obtain other accompanying drawing according to the accompanying drawing provided.
The flow chart of the method for the car plate type approval that Fig. 1 provides for the embodiment of the present invention;
The structured flowchart of the system of the car plate type approval that Fig. 2 provides for the embodiment of the present invention;
The structured flowchart of the system of another car plate type approval that Fig. 3 provides for the embodiment of the present invention.
Detailed description of the invention
The core of the present invention is to provide the method and system of a kind of car plate type approval, it is possible to verify the type of car plate fast and accurately, to some extent solves the problem causing Car license recognition mistake in polymorphic type Car license recognition because type identification is inaccurate.
For making the purpose of the embodiment of the present invention, technical scheme and advantage clearly, below in conjunction with the accompanying drawing in the embodiment of the present invention, technical scheme in the embodiment of the present invention is clearly and completely described, obviously, described embodiment is a part of embodiment of the present invention, rather than whole embodiments. Based on the embodiment in the present invention, the every other embodiment that those of ordinary skill in the art obtain under not making creative work premise, broadly fall into the scope of protection of the invention.
Refer to the flow chart of the method for the car plate type approval that Fig. 1, Fig. 1 provide (3 figure having been carried out literal simple modification, refer to) for the embodiment of the present invention; The method may include that
S100, acquisition license plate image;
Wherein, can be the license plate image obtained from vehicle image here; Can also be the license plate image first obtained from vehicle image, then license plate image be obtained car plate coarse positioning image based on rim detection, then pass through SVM classifier and filter the non-car plate detection block in rough sort, it is thus achieved that License Plate image. The concrete mode obtaining license plate image is not defined by the application, also the quality of the license plate image obtained is not defined; These can select according to user's practical situation, for instance user's calculating computing capability to equipment, and considering factors such as the accuracys rate of result.
S110, by car plate type identification grader, described license plate image is identified, determines the rough sort type of described license plate image according to recognition result according to predetermined condition;
Wherein, the acquisition process of car plate type identification grader here can be through SVM (SupportVectorMachine, support vector machine) the car plate type identification grader trained, it is also possible to is that the training of other algorithms most in use obtains.
Determine the rough sort type of described license plate image according to predetermined condition according to recognition result; Here predetermined condition is the condition selecting rough sort type number; Can be such as that the car plate type of the confidence level front two obtained after SVM identifies is as rough sort type; Can also be as rough sort type according to the several car plate type of the indefinite selection of certain algorithm, such as according to the confidence level obtained after identifying, using the recognition result that meets certain confidence interval as rough sort type, or can also as rough sort type using the car plate type after the identification corresponding to the value of the confidence level in preset range of the difference with most confidence value. Therefore, predetermined condition is not defined here, it is possible to specifically determine according to the demand of user.
S120, call the characters on license plate template corresponding with described rough sort type the car plate in described license plate image is carried out character cutting;
Wherein, herein for the accuracy improving character cutting, call the characters on license plate template corresponding with rough sort type and the car plate in described license plate image is carried out character cutting; Owing to characters on license plate template here is corresponding with car plate type, the characters on license plate template for such car plate being carried out character cutting namely obtained according to the priori of every class car plate.
Optionally, call the characters on license plate template corresponding with described rough sort type the car plate in described license plate image is carried out character cutting to may include that
When described rough sort type is black board, then call blue-black board Character mother plate and car plate is carried out character cutting by embassy's board Character mother plate;
Wherein, black board here can be common black board or embassy's board; Therefore, use blue-black board Character mother plate and embassy's board Character mother plate that car plate is carried out character cutting.
When described rough sort type is monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate or alert board, then call the monolayer yellow card Character mother plate of correspondence, monolayer army board Character mother plate, monolayer People's Armed Police's characters on license plate template or alert board Character mother plate and car plate is carried out character cutting;
Wherein, when described rough sort type is monolayer yellow card, then use monolayer yellow card Character mother plate that this car plate is carried out character cutting; When described rough sort type is monolayer army board, then use monolayer army board Character mother plate that this car plate is carried out character cutting, namely call the characters on license plate template corresponding with car plate type and carry out character cutting.
When described rough sort type is double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card, then call the double-deck army board Character mother plate of correspondence, double-deck People's Armed Police's characters on license plate template or double-deck yellow card Character mother plate and car plate is carried out character cutting;
When described rough sort type is blue board, then calls blue-black board Character mother plate and car plate is carried out character cutting.
Said process is exemplified below:
Result based on car plate type identification uses corresponding board Character mother plate to carry out Character segmentation, concrete disposition is as follows: (1) is if the highest recognition result of confidence level is black board, then directly use blue-black board Character mother plate (common blue-black board is that same template is except embassy's board) and embassy's board Character mother plate that car plate is carried out character cutting; (2) if the highest recognition result of confidence level is the one in monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate and alert board, then use above-mentioned four class characters on license plate templates that car plate is carried out character cutting respectively; (3) if the highest recognition result of confidence level is double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card, then use above-mentioned three class characters on license plate templates that car plate is carried out character cutting respectively; (4) if the highest recognition result of confidence level is blue board, owing to the type identification rate of blue board is higher and blue board is absent from other interference class car plates, follow-up without carrying out independent type approval. Directly use blue-black board Character mother plate to carry out cutting, then carry out the identification of characters on license plate.
S130, according to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding;
Wherein, owing to each type of car plate has its feature, for instance, embassy's board only can contain " making " word, only People's Armed Police's car plate can contain " WJ ", and only alert board can contain " police " word etc.; Therefore, difference character according to each type car plate and the positional information of this difference character, these information pointers are utilized to be trained obtaining character model to the car plate with particularity type, therefore, utilize pointed character model that the character of described car plate differential position is identified, the accuracy of rough sort type can be determined accurately, in order to finally determine the real type of car plate.
The training of character model can use a lot of training algorithm, concrete training method and training algorithm are not defined by the application, as long as can training and obtaining character model, as the accuracy of character model, user can select according to the practical situation of self; Wherein preferred, described character model may include that
Utilize the two LDA model classified and the SVM models that " making " trains;
Utilize the two LDA model classified and the SVM models that " WJ " trains;
Utilize the two LDA model classified and the SVM models that " police " trains;
Utilize the 12 LDA model classified and the SVM models that " 12 army's board initials " trains.
The training process of each character model may is that
First, obtaining the character sample of differential position in all kinds of car plates after cutting by Car license recognition program, the number of character sample can be determined voluntarily; " making " printed words of such as embassy's board this, " WJ " character sample of 12 army's board initial character samples such as " G ", " H " of army's board, People's Armed Police's car plate, " police " character sample etc. of alert board.
Then, all kinds of character samples are carried out data process; Can be strengthen character picture contrast, by enhanced image scaling to preliminary dimension; Such as, strengthen the contrast of grayscale character image, all character pictures are zoomed to wide be 16, height be 32 sizes.
Finally, for embassy's car plate, use " making " character and non-" making " character to be trained, obtain LDA model and the SVM model of corresponding two classification; For army's board, use 12 class characters to be trained, obtain corresponding 12 class LDA models and SVM model; For People's Armed Police's car plate, use " WJ " and non-" WJ " character to be trained, obtain corresponding two classification LDA model and SVM models; For alert board, use " police " character sample and non-" police " character sample to be trained, obtain corresponding two classification LDA model and SVM models.
For three kinds of situations in s120, character recognition process can be (1) owing to common black board initial character Chinese character is province character, and embassy's board is " making " word, so directly using the LDA model of the Chinese character " making " trained and SVM model that initial character is identified, obtain LDA and the SVM confidence level identified, LDA and SVM model herein is all two disaggregated models, namely training sample only comprises two classes, one class is the sample of " making " word, another kind of other samples for non-" making " word, (2) owing to differential position is inconsistent, therefore need to use multiple LDA model and SVM model to be identified checking, character on all kinds of car plate differential position is respectively as follows: the initial character " WJ " (in processes using " WJ " of People's Armed Police's car plate as a character) of monolayer People's Armed Police's car plate, the initial character (being all capitalization) of monolayer army board, the trailing character Chinese character " police " of alert board, owing to generally yellow card car is higher relative to the probability that other special car plates occur, yellow card is not carried out associated verification by this embodiment, if the checking of other types characters on license plate is not passed through, so this car plate type is set to yellow card. LDA model and SVM model that all kinds of car plate type approval use are as follows respectively: for monolayer People's Armed Police's car plate, adopt LDA model and the SVM model of " WJ " character, this model is two disaggregated models, training sample only comprises two classes, one class is " WJ " character sample, another kind of other samples for non-" WJ ". for monolayer army board car plate, employing is the identification checking of initial, initial totally 12, and LDA model and SVM model are based on many disaggregated models of this 12 class-letter sample, for alert board, employing is the strategy the same with monolayer army board, for yellow card, not being verified, above-mentioned three kinds of checkings are not passed through, then the end product of type identification is yellow card. (3) owing to double-deck army board, double-deck People's Armed Police's car plate, the double-deck character of yellow card differential position are consistent with the monolayer car plate of this three class, therefore for this kind of situation, the processing method adopted is consistent with the processing method in step (2), does not repeat them here.
S140, the threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the type of described license plate image.
Wherein, the threshold value of the recognition result obtained Yu character model is compared, according to comparative result, it is determined that the type of license plate image; Here comparative result is that the accuracy requirement according to user is determined. The order of Car license recognition can be made regulation to improve recognition speed, such as, in (2), the checking order of four class car plates is: monolayer People's Armed Police's car plate, monolayer army board, alert board, yellow card (without checking), and the checking order of (3) three class car plates is: double-deck People's Armed Police's car plate, double-deck army board, double-deck yellow card (without checking).
Optionally, the threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the accurate type of described license plate image may include that
LDA model that character according to described car plate differential position is corresponding and SVM model, the LDA model confidence of output and SVM model confidence;
The LDA model threshold that described LDA model confidence is corresponding with the character of described car plate differential position is compared, if more than LDA model threshold, then obtains the first result;
The SVM model threshold that described SVM model confidence is corresponding with the character of described car plate differential position is compared, if more than SVM model threshold, then obtains the second result;
When meeting described first result and described second result simultaneously, then the type of described license plate image is the type that described LDA model is corresponding with SVM model.
Such as: for three kinds of situations in s130, the determination process of car plate type may is that
(1) if the confidence level C1 of LDA output is more than threshold value TH1, and the confidence level C2 of SVM output is more than threshold value TH2, then thinking that embassy's board is verified, car plate type is embassy's board, and otherwise car plate type is common black board;
The checking order of (2) four class car plates is: monolayer People's Armed Police's car plate, monolayer army board, alert board, yellow card (without checking), the mode of checking is consistent with step (1), also it is the checking that respectively the differential position character of each class car plate is carried out LDA and SVM confidence level, threshold value only for each class characters on license plate can be variant, therefore does not repeat them here.
The checking order of (3) three class car plates is: double-deck People's Armed Police's car plate, double-deck army board, double-deck yellow card (without checking), the mode of checking is consistent with step (1), does not repeat them here.
Through said process, the rough sort type obtained further is verified, improve the accuracy of car plate type identification, and then solve that type identification in polymorphic type Car license recognition is inaccurate and the problem that causes Car license recognition mistake.
Based on technique scheme, the method of the car plate type approval that the embodiment of the present invention provides, by using car plate type identification grader that the type of car plate is classified, utilize the characters on license plate template corresponding with classification results that the character in car plate is carried out cutting, the character of car plate differential position is identified by the character model utilizing the character of car plate differential position corresponding, and determine that whether the type that car plate type identification grader obtains is accurate according to recognition result, and finally determine the accurate type of car plate according to recognition result; Namely the method is after the type of car plate is classified by car plate type identification grader, and sorted result is verified again; Therefore, it is possible to verify the type of car plate fast and accurately, to some extent solve the problem causing Car license recognition mistake in polymorphic type Car license recognition because type identification is inaccurate.
Based on technique scheme, the method can also include:
Regularly described character model is safeguarded.
Wherein, the accuracy that rough sort type is verified in process is needed to rely on the accuracy of character model by the method, accordingly, it would be desirable to regularly described character model is safeguarded, thus ensureing the accuracy of character model identification. Such as can by according to actual identification error rate decide whether increase sample size character model is updated, or whether produce new outstanding computational methods character model is carried out safeguard update.
Based on technique scheme, the method of the car plate type approval that the embodiment of the present invention provides, multiple car plate type identification first passes through car plate type identification grader car plate type is slightly identified, then the characters on license plate template using respective type carries out license plate character segmentation, all kinds of car plates have the character of differential position use LDA and SVM training in advance respectively good character model is identified, carry out Authentication-Type by the confidence level of LDA and SVM. One character of some class car plate is identified checking by this kind of method, if by verifying, follow-up without repeating to identify this position character, the type of car plate can be verified by this kind of method fast and accurately, to some extent solve that type identification in polymorphic type Car license recognition is inaccurate and the problem that causes Car license recognition mistake.
The method embodiments providing car plate type approval, can verify the type of car plate fast and accurately by said method.
Below the system of the car plate type approval that the embodiment of the present invention provides being introduced, the method for the system of car plate type approval described below and above-described car plate type approval can mutually to should refer to.
Refer to the structured flowchart of the system of the car plate type approval that Fig. 2, Fig. 2 provide for the embodiment of the present invention; This system may include that
Acquisition module 100, is used for obtaining license plate image;
Rough sort module 200, for described license plate image being identified by car plate type identification grader, determines the rough sort type of described license plate image according to recognition result according to predetermined condition;
Cutting module 300, carries out character cutting for calling the characters on license plate template corresponding with described rough sort type to the car plate in described license plate image;
Character recognition module 400, for according to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding;
Determination type module 500, for comparing the threshold value of character model corresponding with the character of described car plate differential position for recognition result, it is determined that the type of described license plate image.
Optionally, described cutting module 300 includes:
First cutting unit, for being black board when described rough sort type, then calls blue-black board Character mother plate and car plate is carried out character cutting by embassy's board Character mother plate;
Second cutting unit, for being monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate or alert board when described rough sort type, then call the monolayer yellow card Character mother plate of correspondence, monolayer army board Character mother plate, monolayer People's Armed Police's characters on license plate template or alert board Character mother plate and car plate is carried out character cutting;
3rd cutting unit, for being double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card when described rough sort type, then calls the double-deck army board Character mother plate of correspondence, double-deck People's Armed Police's characters on license plate template or double-deck yellow card Character mother plate and car plate is carried out character cutting;
4th cutting unit, for being blue board when described rough sort type, then calls blue-black board Character mother plate and car plate is carried out character cutting.
Optionally, this system may include that
First training module, is used for the two LDA model classified and the SVM models utilizing " making " to train;
Second training module, is used for the two LDA model classified and the SVM models utilizing " WJ " to train;
3rd training module, is used for the two LDA model classified and the SVM models utilizing " police " to train;
4th training module, is used for the 12 LDA model classified and the SVM models utilizing " 12 army's board initials " to train.
Optionally, described determination type module 500 includes:
Output unit, for the LDA model corresponding according to the character of described car plate differential position and SVM model, the LDA model confidence of output and SVM model confidence;
First comparing unit, for the LDA model threshold that described LDA model confidence is corresponding with the character of described car plate differential position being compared, if more than LDA model threshold, then obtains the first result;
Second comparing unit, for the SVM model threshold that described SVM model confidence is corresponding with the character of described car plate differential position being compared, if more than SVM model threshold, then obtains the second result;
Type determining units, for when meeting described first result and described second result simultaneously, then the type of described license plate image is the type that described LDA model is corresponding with SVM model.
Based on technique scheme, refer to Fig. 3, this system can also include:
Maintenance module 600, for regularly safeguarding described character model.
In description, each embodiment adopts the mode gone forward one by one to describe, and what each embodiment stressed is the difference with other embodiments, between each embodiment identical similar portion mutually referring to. For device disclosed in embodiment, owing to it corresponds to the method disclosed in Example, so what describe is fairly simple, relevant part illustrates referring to method part.
Professional further appreciates that, the unit of each example described in conjunction with the embodiments described herein and algorithm steps, can with electronic hardware, computer software or the two be implemented in combination in, in order to clearly demonstrate the interchangeability of hardware and software, generally describe composition and the step of each example in the above description according to function. These functions perform with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme. Professional and technical personnel specifically can should be used for using different methods to realize described function to each, but this realization is it is not considered that beyond the scope of this invention.
The method described in conjunction with the embodiments described herein or the step of algorithm can directly use the software module that hardware, processor perform, or the combination of the two is implemented. Software module can be placed in any other form of storage medium known in random access memory (RAM), internal memory, read only memory (ROM), electrically programmable ROM, electrically erasable ROM, depositor, hard disk, moveable magnetic disc, CD-ROM or technical field.
Above the method and system of car plate type approval provided by the present invention are described in detail. Principles of the invention and embodiment are set forth by specific case used herein, and the explanation of above example is only intended to help to understand method and the core concept thereof of the present invention. It should be pointed out that, for those skilled in the art, under the premise without departing from the principles of the invention, it is also possible to the present invention carries out some improvement and modification, these improve and modify in the protection domain also falling into the claims in the present invention.
Claims (10)
1. the method for a car plate type approval, it is characterised in that including:
Obtain license plate image;
By car plate type identification grader, described license plate image is identified, determines the rough sort type of described license plate image according to recognition result according to predetermined condition;
Call the characters on license plate template corresponding with described rough sort type and the car plate in described license plate image is carried out character cutting;
According to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding;
The threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the type of described license plate image.
2. the method for claim 1, it is characterised in that call the characters on license plate template corresponding with described rough sort type and the car plate in described license plate image is carried out character cutting, including:
When described rough sort type is black board, then call blue-black board Character mother plate and car plate is carried out character cutting by embassy's board Character mother plate;
When described rough sort type is monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate or alert board, then call the monolayer yellow card Character mother plate of correspondence, monolayer army board Character mother plate, monolayer People's Armed Police's characters on license plate template or alert board Character mother plate and car plate is carried out character cutting;
When described rough sort type is double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card, then call the double-deck army board Character mother plate of correspondence, double-deck People's Armed Police's characters on license plate template or double-deck yellow card Character mother plate and car plate is carried out character cutting;
When described rough sort type is blue board, then calls blue-black board Character mother plate and car plate is carried out character cutting.
3. method as claimed in claim 2, it is characterised in that described character model includes:
Utilize the two LDA model classified and the SVM models that " making " trains;
Utilize the two LDA model classified and the SVM models that " WJ " trains;
Utilize the two LDA model classified and the SVM models that " police " trains;
Utilize the 12 LDA model classified and the SVM models that " 12 army's board initials " trains.
4. method as claimed in claim 3, it is characterised in that the threshold value of character model corresponding with the character of described car plate differential position for recognition result is compared, it is determined that the accurate type of described license plate image, including:
LDA model that character according to described car plate differential position is corresponding and SVM model, the LDA model confidence of output and SVM model confidence;
The LDA model threshold that described LDA model confidence is corresponding with the character of described car plate differential position is compared, if more than LDA model threshold, then obtains the first result;
The SVM model threshold that described SVM model confidence is corresponding with the character of described car plate differential position is compared, if more than SVM model threshold, then obtains the second result;
When meeting described first result and described first result simultaneously, then the type of described license plate image is the type that described LDA model is corresponding with SVM model.
5. the method as described in any one of Claims 1-4, it is characterised in that also include:
Regularly described character model is safeguarded.
6. the system of a car plate type approval, it is characterised in that including:
Acquisition module, is used for obtaining license plate image;
Rough sort module, for described license plate image being identified by car plate type identification grader, determines the rough sort type of described license plate image according to recognition result according to predetermined condition;
Cutting module, carries out character cutting for calling the characters on license plate template corresponding with described rough sort type to the car plate in described license plate image;
Character recognition module, for according to described rough sort type, it is determined that the character of car plate differential position, and the character of described car plate differential position is identified by the character model utilizing the character of described car plate differential position corresponding;
Determination type module, for comparing the threshold value of character model corresponding with the character of described car plate differential position for recognition result, it is determined that the type of described license plate image.
7. system as claimed in claim 6, it is characterised in that described cutting module includes:
First cutting unit, for being black board when described rough sort type, then calls blue-black board Character mother plate and car plate is carried out character cutting by embassy's board Character mother plate;
Second cutting unit, for being monolayer yellow card, monolayer army board, monolayer People's Armed Police's car plate or alert board when described rough sort type, then call the monolayer yellow card Character mother plate of correspondence, monolayer army board Character mother plate, monolayer People's Armed Police's characters on license plate template or alert board Character mother plate and car plate is carried out character cutting;
3rd cutting unit, for being double-deck army board, double-deck People's Armed Police's car plate or double-deck yellow card when described rough sort type, then calls the double-deck army board Character mother plate of correspondence, double-deck People's Armed Police's characters on license plate template or double-deck yellow card Character mother plate and car plate is carried out character cutting;
4th cutting unit, for being blue board when described rough sort type, then calls blue-black board Character mother plate and car plate is carried out character cutting.
8. system as claimed in claim 7, it is characterised in that including:
First training module, is used for the two LDA model classified and the SVM models utilizing " making " to train;
Second training module, is used for the two LDA model classified and the SVM models utilizing " WJ " to train;
3rd training module, is used for the two LDA model classified and the SVM models utilizing " police " to train;
4th training module, is used for the 12 LDA model classified and the SVM models utilizing " 12 army's board initials " to train.
9. system as claimed in claim 8, it is characterised in that described determination type module includes:
Output unit, for the LDA model corresponding according to the character of described car plate differential position and SVM model, the LDA model confidence of output and SVM model confidence;
First comparing unit, for the LDA model threshold that described LDA model confidence is corresponding with the character of described car plate differential position being compared, if more than LDA model threshold, then obtains the first result;
Second comparing unit, for the SVM model threshold that described SVM model confidence is corresponding with the character of described car plate differential position being compared, if more than SVM model threshold, then obtains the second result;
Type determining units, for when meeting described first result and described first result simultaneously, then the type of described license plate image is the type that described LDA model is corresponding with SVM model.
10. the system as described in any one of claim 6 to 9, it is characterised in that also include:
Maintenance module, for regularly safeguarding described character model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510969964.1A CN105631470A (en) | 2015-12-21 | 2015-12-21 | Method and system for verifying license plate type |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510969964.1A CN105631470A (en) | 2015-12-21 | 2015-12-21 | Method and system for verifying license plate type |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105631470A true CN105631470A (en) | 2016-06-01 |
Family
ID=56046382
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510969964.1A Pending CN105631470A (en) | 2015-12-21 | 2015-12-21 | Method and system for verifying license plate type |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105631470A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106326897A (en) * | 2016-08-24 | 2017-01-11 | 王小兰 | License plate number recognition method based on data model |
| WO2017220032A1 (en) * | 2016-06-24 | 2017-12-28 | 平安科技（深圳）有限公司 | Vehicle license plate classification method and system based on deep learning, electronic apparatus, and storage medium |
| WO2018090771A1 (en) * | 2016-11-16 | 2018-05-24 | 杭州海康威视数字技术股份有限公司 | Vehicle license plate recognition method and apparatus |
| CN108073926A (en) * | 2016-11-17 | 2018-05-25 | 杭州海康威视数字技术股份有限公司 | A kind of licence plate recognition method and device |
| CN108090484A (en) * | 2016-11-23 | 2018-05-29 | 杭州海康威视数字技术股份有限公司 | A kind of licence plate recognition method and device |
| CN108229466A (en) * | 2016-12-15 | 2018-06-29 | 杭州海康威视数字技术股份有限公司 | A kind of licence plate recognition method and device |
| CN109063702A (en) * | 2018-08-08 | 2018-12-21 | 北京旷视科技有限公司 | Licence plate recognition method, device, equipment and storage medium |
| CN109657664A (en) * | 2017-10-12 | 2019-04-19 | 杭州海康威视数字技术股份有限公司 | A kind of recognition methods, device and the electronic equipment of license plate type |
| CN109726678A (en) * | 2018-12-28 | 2019-05-07 | 深圳市捷顺科技实业股份有限公司 | A kind of method and relevant apparatus of Car license recognition |
| CN110009058A (en) * | 2019-04-16 | 2019-07-12 | 山东博昂信息科技有限公司 | A kind of parking lot Vehicle License Plate Recognition System and method |
| CN110533039A (en) * | 2019-09-04 | 2019-12-03 | 深圳市捷顺科技实业股份有限公司 | A kind of true-false detection method of license plate, device and equipment |
| CN110728278A (en) * | 2018-07-16 | 2020-01-24 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device |
| CN111582261A (en) * | 2020-04-30 | 2020-08-25 | 浙江大华技术股份有限公司 | License plate recognition method and device for non-motor vehicle |
| CN112598005A (en) * | 2020-12-24 | 2021-04-02 | 深圳市信义科技有限公司 | Character segmentation method for license plate of armed police, storage medium and computer equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6026177A (en) * | 1995-08-29 | 2000-02-15 | The Hong Kong University Of Science & Technology | Method for identifying a sequence of alphanumeric characters |
| CN101673338A (en) * | 2009-10-09 | 2010-03-17 | 南京树声科技有限公司 | Fuzzy license plate identification method based on multi-angle projection |
| CN102722733A (en) * | 2012-05-31 | 2012-10-10 | 信帧电子技术（北京）有限公司 | Identification method and device of license plate types |
| CN104268596A (en) * | 2014-09-25 | 2015-01-07 | 深圳市捷顺科技实业股份有限公司 | License plate recognizer and license plate detection method and system thereof |
| CN104392239A (en) * | 2014-09-25 | 2015-03-04 | 深圳市捷顺科技实业股份有限公司 | License plate identification method and system |
-
2015
- 2015-12-21 CN CN201510969964.1A patent/CN105631470A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6026177A (en) * | 1995-08-29 | 2000-02-15 | The Hong Kong University Of Science & Technology | Method for identifying a sequence of alphanumeric characters |
| CN101673338A (en) * | 2009-10-09 | 2010-03-17 | 南京树声科技有限公司 | Fuzzy license plate identification method based on multi-angle projection |
| CN102722733A (en) * | 2012-05-31 | 2012-10-10 | 信帧电子技术（北京）有限公司 | Identification method and device of license plate types |
| CN104268596A (en) * | 2014-09-25 | 2015-01-07 | 深圳市捷顺科技实业股份有限公司 | License plate recognizer and license plate detection method and system thereof |
| CN104392239A (en) * | 2014-09-25 | 2015-03-04 | 深圳市捷顺科技实业股份有限公司 | License plate identification method and system |
Non-Patent Citations (1)
| Title |
|---|
| DERDOUR KHEDIDJA等: "Digit recognition using multiple classifiers", 《PROGRAMMING AND SYSTEMS (ISPS), 2015 12TH INTERNATIONAL SYMPOSIUM ON》 * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10528841B2 (en) | 2016-06-24 | 2020-01-07 | Ping An Technology (Shenzhen) Co., Ltd. | Method, system, electronic device, and medium for classifying license plates based on deep learning |
| WO2017220032A1 (en) * | 2016-06-24 | 2017-12-28 | 平安科技（深圳）有限公司 | Vehicle license plate classification method and system based on deep learning, electronic apparatus, and storage medium |
| CN106326897A (en) * | 2016-08-24 | 2017-01-11 | 王小兰 | License plate number recognition method based on data model |
| WO2018090771A1 (en) * | 2016-11-16 | 2018-05-24 | 杭州海康威视数字技术股份有限公司 | Vehicle license plate recognition method and apparatus |
| CN108073926A (en) * | 2016-11-17 | 2018-05-25 | 杭州海康威视数字技术股份有限公司 | A kind of licence plate recognition method and device |
| CN108090484A (en) * | 2016-11-23 | 2018-05-29 | 杭州海康威视数字技术股份有限公司 | A kind of licence plate recognition method and device |
| CN108229466A (en) * | 2016-12-15 | 2018-06-29 | 杭州海康威视数字技术股份有限公司 | A kind of licence plate recognition method and device |
| CN108229466B (en) * | 2016-12-15 | 2020-12-11 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device |
| CN109657664A (en) * | 2017-10-12 | 2019-04-19 | 杭州海康威视数字技术股份有限公司 | A kind of recognition methods, device and the electronic equipment of license plate type |
| CN110728278A (en) * | 2018-07-16 | 2020-01-24 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device |
| CN109063702A (en) * | 2018-08-08 | 2018-12-21 | 北京旷视科技有限公司 | Licence plate recognition method, device, equipment and storage medium |
| CN109726678A (en) * | 2018-12-28 | 2019-05-07 | 深圳市捷顺科技实业股份有限公司 | A kind of method and relevant apparatus of Car license recognition |
| CN109726678B (en) * | 2018-12-28 | 2023-02-28 | 深圳市捷顺科技实业股份有限公司 | License plate recognition method and related device |
| CN110009058A (en) * | 2019-04-16 | 2019-07-12 | 山东博昂信息科技有限公司 | A kind of parking lot Vehicle License Plate Recognition System and method |
| CN110533039A (en) * | 2019-09-04 | 2019-12-03 | 深圳市捷顺科技实业股份有限公司 | A kind of true-false detection method of license plate, device and equipment |
| CN111582261A (en) * | 2020-04-30 | 2020-08-25 | 浙江大华技术股份有限公司 | License plate recognition method and device for non-motor vehicle |
| CN111582261B (en) * | 2020-04-30 | 2024-01-19 | 浙江大华技术股份有限公司 | License plate recognition method and license plate recognition device for non-motor vehicle |
| CN112598005A (en) * | 2020-12-24 | 2021-04-02 | 深圳市信义科技有限公司 | Character segmentation method for license plate of armed police, storage medium and computer equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105631470A (en) | Method and system for verifying license plate type | |
| CN106156766B (en) | Method and device for generating text line classifier | |
| CN106991820B (en) | Illegal vehicle processing method and device | |
| US9418304B2 (en) | System and method for recognizing text information in object | |
| CN103488973B (en) | Vehicle brand recognition methods and system based on image | |
| CN111523464A (en) | Method and device for detecting illegal lane change of vehicle | |
| CN104766042A (en) | Method and apparatus for and recognizing traffic sign board | |
| CN109740424A (en) | Traffic violations recognition methods and Related product | |
| Han et al. | Real‐time license plate detection in high‐resolution videos using fastest available cascade classifier and core patterns | |
| CN109657664A (en) | A kind of recognition methods, device and the electronic equipment of license plate type | |
| CN103914680A (en) | Character image jet-printing, recognition and calibration system and method | |
| CN111242002B (en) | Shared bicycle standardized parking judgment method based on computer vision | |
| CN107529659B (en) | Seatbelt wearing detection method, device and electronic equipment | |
| CN105894819A (en) | Fake-license-plate vehicle identification method based on twice verification | |
| CN104077594A (en) | Image recognition method and device | |
| CN109657674A (en) | A kind of vehicles identifications determine method and device | |
| CN110765963A (en) | Vehicle brake detection method, device, equipment and computer readable storage medium | |
| CN104156701A (en) | Plate number similar character recognition method based on decision-making tree and SVM | |
| CN110619255B (en) | Target detection method and device | |
| Lu et al. | Traffic light recognition | |
| CN106778765B (en) | License plate recognition method and device | |
| CN112329886A (en) | Double-license plate recognition method, model training method, device, equipment and storage medium | |
| CN103065141B (en) | Based on the license plate locating method of Color-based clustering | |
| CN108985197B (en) | Automatic detection method for taxi driver smoking behavior based on multi-algorithm fusion | |
| CN108830182B (en) | Lane line detection method based on cascade convolution neural network |
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: 20160601 |
|
| RJ01 | Rejection of invention patent application after publication |