CN109063702A - Licence plate recognition method, device, equipment and storage medium - Google Patents
Licence plate recognition method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN109063702A CN109063702A CN201810899533.6A CN201810899533A CN109063702A CN 109063702 A CN109063702 A CN 109063702A CN 201810899533 A CN201810899533 A CN 201810899533A CN 109063702 A CN109063702 A CN 109063702A
- Authority
- CN
- China
- Prior art keywords
- license plate
- confidence level
- recognition result
- plate recognition
- maximum
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
- G06V10/23—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on positionally close patterns or neighbourhood relationships
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Character Discrimination (AREA)
- Traffic Control Systems (AREA)
Abstract
Licence plate recognition method, device, equipment and storage medium provided by the invention, belong to technical field of image processing.This method comprises: obtaining the license plate recognition result identified by step-by-step recognition methods；License plate recognition result is assumed correspondingly with the multiple default license plate type according to the processing rule of multiple default license plate types and license plate recognition result are determining；Determine the corresponding confidence level of multiple hypothesis license plate recognition result institutes；Select hypothesis license plate recognition result corresponding to maximum confidence level as target license plate recognition result from multiple confidence levels.License plate recognition result is assumed correspondingly with multiple default license plate types by generating, so that xenogenesis license plate will retain higher confidence level, and the license plate of error type confidence level after through above-mentioned processing will substantially reduce, and then the confidence level by comparing the license plate recognition result of license plate after having passed through the processing rule of different default license plate types, and then effectively increase the accuracy to Car license recognition.
Description
Technical field
The present invention relates to field of image processings, are situated between in particular to licence plate recognition method, device, equipment and storage
Matter.
Background technique
The detection of license plate and identification technology are always a ring important in safety monitoring environment, are such as supervised in automatic parking lot
It controls, in the application scenarios such as high speed bayonet monitoring, there is the demand of license plate in clear identification picture.Therefore, existing Car license recognition
Technology has also obtained systematically developing, among numerous algorithms, be generally used sorting algorithm for license plate each character by
Position is classified, and the character an of prediction and the confidence level to the character is obtained, after every result of license plate is spliced together
Final result is obtained, every confidence level is multiplied to obtain the recognition confidence of license plate.
However there are many defects in the accuracy of identification for existing method, sixty-four dollar question is, in order to simultaneous
Hold some special xenogenesis license plates (such as non-Chinese character starts license plate, the exception Chinese character such as embassy/police car license plate) of identification and passes through the method
It can identify forbidden character in some locations, such as first place of the common license plate of blue of license plate is not Chinese character, or blue is commonly
Situations such as second of license plate is not letter, or 0 the problem of obscuring of letter D and number.
Summary of the invention
Licence plate recognition method, device, equipment and storage medium provided in an embodiment of the present invention, can solve the prior art
In Car license recognition the lower technical problem of precision.
To achieve the goals above, technical solution used in the embodiment of the present invention is as follows:
In a first aspect, a kind of licence plate recognition method provided in an embodiment of the present invention, comprising: acquisition passes through step-by-step identification side
The license plate recognition result that method identifies；It is determined according to the processing rule of multiple default license plate types and the license plate recognition result
License plate recognition result is assumed correspondingly with the multiple default license plate type；Determine multiple hypothesis Car license recognition knots
The corresponding confidence level of fruit institute；The hypothesis corresponding to the maximum confidence level is selected from multiple confidence levels
License plate recognition result is as target license plate recognition result.
With reference to first aspect, described the embodiment of the invention provides the first possible embodiment of first aspect
According to the processing rule of multiple default license plate types and license plate recognition result determination and the multiple default license plate type one
One corresponding hypothesis license plate recognition result, comprising: the processing of the multiple default license plate type is determined according to license plate recognition result
The hypothesis position of regular institute license plate correspondingly；Determine the hypothesis position and the word for assuming the pervious all character bits in position
The product of the corresponding confidence level of symbol；Determine character string corresponding to the maximum value in the product；Concatenated according to the character
License plate recognition result is assumed correspondingly at the multiple default license plate type.
With reference to first aspect, described the embodiment of the invention provides second of possible embodiment of first aspect
Select the hypothesis license plate recognition result corresponding to the maximum confidence level as target carriage from multiple confidence levels
Board recognition result, comprising: determine each hypothesis Car license recognition for the processing rule of default license plate type described in each
As a result medium size code bit includes the total confidence level of maximum of most two English alphabets；According to maximum total confidence level determination
Post-processing recognition result corresponding to number position；From the regular corresponding multiple institutes of processing with multiple default license plate types
Stating in post-processing recognition result selects confidence level maximum as target license plate recognition result.
The possible embodiment of second with reference to first aspect, the embodiment of the invention provides the thirds of first aspect
The possible embodiment of kind, the processing rule for each default license plate type determine the target license plate
Recognition result medium size code bit includes the total confidence level of maximum of most two English alphabets, comprising: is directed to each described default vehicle
The processing rule of board type determines in number position to include first total confidence level of zero-bit English alphabet respectively, includes an English
Second alphabetical total confidence level and the total confidence level of third comprising two English alphabets；By described first total confidence level, described
Numerical value is maximum as maximum total confidence level in second total confidence level and the total confidence level of the third.
The third possible embodiment with reference to first aspect, the embodiment of the invention provides the 4th of first aspect the
The possible embodiment of kind, the number position include N characters, and the N is the integer more than or equal to 4, and described being directed to is every
The processing rule of one default license plate type determines first total confidence in number position comprising zero-bit English alphabet respectively
Degree, second comprising an English alphabet total confidence level and the total confidence level of third comprising two English alphabets, comprising: needle
Confidence is obtained from confidence level corresponding to multiple preset numbers respectively the processing rule of each default license plate type
It spends maximum digital confidence level and obtains the maximum letter of confidence level from confidence level corresponding to multiple default letters and set
Reliability；First product of the first confidence level and the digital confidence level before obtaining in M comprising J English alphabets, the M
Subtract one equal to the N, the J is the integer less than or equal to 2；Obtain includes the second of X English alphabets in described preceding M
Second product of confidence level and the alphabetical confidence level, the X are equal to the J and subtract one；When J is 0, by first product
It is maximum as first total confidence level with numerical value in the second product；It, will be in second product of the first sum of products when J is 1
Numerical value is maximum to be used as second total confidence level；When J is 2, by the maximum conduct of numerical value in second product of the first sum of products
The total confidence level of third.
The possible embodiment of second with reference to first aspect, the embodiment of the invention provides the 5th of first aspect the
Kind possible embodiment, it is described from the processing rule with multiple default license plate types it is corresponding it is multiple it is described after from
Reason recognition result in select confidence level maximum as target license plate recognition result, comprising: from multiple default license plate kinds
It is determined corresponding to each post-processing recognition result in the corresponding multiple post-processing recognition results of processing rule of class
Confidence level；The maximum post-processing recognition result of confidence level is selected to identify from multiple confidence levels as target license plate
As a result.
Second aspect, a kind of license plate recognition device provided in an embodiment of the present invention, comprising: module is obtained, it is logical for obtaining
Cross the license plate recognition result that step-by-step recognition methods identifies；First data processing module, for according to multiple default license plate types
Processing rule and the license plate recognition result is determining assumes Car license recognition correspondingly with the multiple default license plate type
As a result；Second data processing module, for determining the corresponding confidence level of multiple hypothesis license plate recognition result institutes；The
Three data processing modules, for selecting the hypothesis vehicle corresponding to the maximum confidence level from multiple confidence levels
Board recognition result is as target license plate recognition result.
In conjunction with second aspect, the embodiment of the invention provides the first possible embodiment of second aspect, described
Three data processing modules are also used to: determining each hypothesis vehicle for the processing rule of default license plate type described in each
Board recognition result medium size code bit includes the total confidence level of maximum of most two English alphabets；It is true according to maximum total confidence level
Post-processing recognition result corresponding to the fixed number position；From corresponding with the processing rule of multiple default license plate types
Select confidence level maximum as target license plate recognition result in multiple post-processing recognition results.
The third aspect, a kind of terminal device provided in an embodiment of the present invention, comprising: memory, processor and be stored in
In the memory and the computer program that can run on the processor, the processor execute the computer program
Shi Shixian is as described in any one of first aspect the step of licence plate recognition method.
Fourth aspect, a kind of storage medium provided in an embodiment of the present invention are stored with instruction on the storage medium, when
When described instruction is run on computers, so that the computer executes such as the described in any item Car license recognition sides of first aspect
Method.
Compared with prior art, the embodiment of the present invention bring it is following the utility model has the advantages that
Licence plate recognition method, device, equipment and storage medium provided in an embodiment of the present invention, by generate with it is described more
A default license plate type assumes license plate recognition result correspondingly, so that xenogenesis license plate will retain higher confidence
Degree, and the license plate of error type confidence level after through above-mentioned processing will substantially reduce, and then know by comparing the license plate of license plate
Confidence level of the other result after having passed through the processing rule of different default license plate types, and then effectively increase and license plate is known
Other accuracy, to obtain the higher license plate recognition result of accuracy, overcome in the presence of the prior art to Car license recognition
The low technical problem of accuracy.
Other feature and advantage of the disclosure will illustrate in the following description, alternatively, Partial Feature and advantage can be with
Deduce from specification or unambiguously determine, or by implement the disclosure above-mentioned technology it can be learnt that.
To enable the above objects, features and advantages of the present invention to be clearer and more comprehensible, preferred embodiment is cited below particularly, and match
Appended attached drawing is closed, is described in detail below.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be to needed in the embodiment
Attached drawing is briefly described, it should be understood that the following drawings illustrates only certain embodiments of the present invention, therefore is not to be seen as
It is the restriction to range, it for those of ordinary skill in the art, without creative efforts, can be with
Other relevant attached drawings are obtained according to these attached drawings.
Fig. 1 is the flow chart for the licence plate recognition method that first embodiment of the invention provides；
Fig. 2 is that one of licence plate recognition method shown in FIG. 1 identifies structure flow chart；
Fig. 3 is another identification structure flow chart in licence plate recognition method shown in FIG. 1；
Fig. 4 is the functional block diagram for the license plate recognition device that second embodiment of the invention provides；
Fig. 5 is a kind of schematic diagram for terminal device that third embodiment of the invention provides.
Specific embodiment
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, ordinary skill people
Member's every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.Cause
This, is not intended to limit claimed invention to the detailed description of the embodiment of the present invention provided in the accompanying drawings below
Range, but be merely representative of selected embodiment of the invention.Based on the embodiments of the present invention, ordinary skill people
Member's every other embodiment obtained without creative efforts, shall fall within the protection scope of the present invention.
With reference to the accompanying drawing, it elaborates to some embodiments of the present invention.In the absence of conflict, following
Embodiment and embodiment in feature can be combined with each other.
First embodiment
Since there are many defects in the recognition accuracy of existing licence plate recognition method, in order to make up the defect to mention
The precision of height identification, the present embodiment provide firstly a kind of licence plate recognition method, it should be noted that in the flow chart of attached drawing
The step of showing can execute in a computer system such as a set of computer executable instructions, although also, in flow chart
In show logical order, but in some cases, shown or described step can be executed with the sequence for being different from herein
Suddenly.It describes in detail below to the present embodiment.
Referring to Fig. 1, being the flow chart of licence plate recognition method provided in an embodiment of the present invention.It below will be to shown in FIG. 1
Detailed process is described in detail.
Step S101 obtains the license plate recognition result identified by step-by-step recognition methods.
License plate recognition result includes multiple character bits.
Optionally, license plate recognition result includes 9 character bits, and every character bit corresponds to a character identification result, each
Character identification result is made of a real number of character set length (digital 0-9, the Chinese character or English alphabet that can occur in license plate),
I-th of real number represents the probability (i.e. confidence level) that the character may be i-th kind of character.Wherein, 9 character identification results include 8
Position license plate number and a null character, null character position are used when increasing for later period license plate digit.
In actual use, the digit of license plate recognition result can be adjusted according to the license plate rule of country variant,
Here, being not especially limited.
As an example it is assumed that in character set corresponding to first character identification result in license plate recognition result
One real number is 0.9 and corresponding character is A, and second real number is 0.8 and corresponding character is B, then first real number 0.9
Represent the probability that first character identification result may be character A.
Optionally, the license plate recognition result can be the matrix of K*L, wherein K is used to characterize the digit of character bit, L
For characterizing character set length corresponding to the corresponding character identification result of every character bit.For example, the license plate recognition result
It can be the matrix of 9*74.
In actual use, license plate recognition result be generally used sorting algorithm for license plate each character step-by-step into
Row classification obtains the character of L prediction for each character bit and to a confidence level of each prediction character., by license plate
The position K character identification result be spliced together after obtained final result (as license plate recognition result).For example, sorting algorithm
It can be but not limited to Bayes Method, decision tree or support vector machines etc..
Step S102, according to the processing rule of multiple default license plate types and the license plate recognition result it is determining with it is described
Multiple default license plate types assume license plate recognition result correspondingly.
The processing rule of multiple default license plate types include common seven license plates processing rule, police car license plate processing rule,
Embassy's license plate processing rule, consulate's license plate handle rule, government's license plate handles rule, the processing of agricultural license plate is regular, small new
Energy license plate processing rule and big new energy license plate processing rule.
In order to more intuitively embody the processing rule of multiple default license plate types, as shown in Table 1:
Table one
As a kind of possible implementation, step S102 includes: to determine the multiple preset according to license plate recognition result
License plate type processing rule institute correspondingly license plate hypothesis position；Determine that the hypothesis position and the hypothesis position are pervious
The product of confidence level corresponding to the character of all character bits；Determine character string corresponding to the maximum value in the product；Root
License plate recognition result is assumed correspondingly according to the text string generation and the multiple default license plate type.
In which it is assumed that position refers to a certain position in license plate recognition result in 9 or two.It such as can be Car license recognition knot
The first place of fruit or last bit or be intermediate any position.For example, it is assumed that license plate recognition result is alert for capital P72U6, if false
It is positioned as last bit, then assumes the corresponding character in position for police.
For example with the processing rule of one of license plate type, for example, police car license plate processing rule, it is assumed that license plate A
Alert for capital P72U6, vacation is positioned as last bit, it is assumed that first corresponding L confidence level is respectively a1, a2 ... aL, second pair
The L confidence level answered is respectively b1, b2 ... bL, and the corresponding L confidence level in third position is respectively c1, c2 ... cL, the 4th correspondence
L confidence level be respectively d1, d2 ... dL, the 5th corresponding L confidence level is respectively e1, e2 ... eL, and the 6th is corresponding
L confidence level is respectively f1, f2 ... fL, as it is assumed that position is the 7th, therefore the 7th corresponding confidence level is g1, then described
It is assumed that the product of confidence level corresponding to position and the character for assuming the pervious all character bits in position is respectively on each
The product of the confidence level of any one character in the confidence level of any one character and other each positions, it may be assumed that CJ1=a1*
B1*c1*d1*e1*f1*g1；CJ2=a2*b2*c2*d2*e2*f2*g1；CJ3=a2*b1*c2*d2*
E2*f2*g1；CJ4=a2*b1*c1*d2*e2*f2*g1 etc. the case where according to various combinations, has A product, from L6The maximum product of product is selected in a product
Corresponding character string.Assuming that the value of product CJ1 is maximum value, and the corresponding character string of CJ1 is ' capital P72U6 is alert ', then
The hypothesis license plate recognition result corresponding with police car license plate processing rule arrived is ' capital P72U6 is alert '.
The processing rule of license plate type default for every kind assumes that character corresponding to position, should for the rule
Confidence level corresponding to character is maximum.For example, last bit is right for the character of ' police ' for the processing of police car license plate rule
The opposite confidence level corresponding in other character set in last bit of the confidence level answered is maximum.For example, facing a Car license recognition
As a result, it is assumed that the license plate is certain license plate, then adds such license plate to this license plate recognition result and thinks to limit, for example, such as
Fruit thinks that a license plate is police car license plate, then can specify that the last bit of the license plate must be ' police ' character.
In the present embodiment, Car license recognition knot is assumed correspondingly with the multiple default license plate type by generating
Fruit, so that xenogenesis license plate will retain higher confidence level, and the license plate of error type confidence after through above-mentioned processing
Degree will substantially reduce, and then by comparing the license plate recognition result of license plate in the processing for having passed through different default license plate types
Confidence level after rule, to obtain the higher license plate recognition result of accuracy.
Step S103 determines the corresponding confidence level of multiple hypothesis license plate recognition result institutes.
Due to having obtained a hypothesis license plate recognition result by the processing rule of every kind of default license plate type, and should
It is assumed that license plate recognition result is corresponding with a confidence level.
As an example it is assumed that the license plate of a vehicle is capital P72U6D, then in the processing rule by presetting license plate type
The recognition result that the processing rule of common seven license plates obtains may be ' capital P72U6D ', confidence level 0.96；And by pre-
If being set to character due to securing last position one when the processing rule of the police car license plate in the processing rule of license plate type
' police ', i.e. vacation are positioned as ' warning ', then recognition result should be ' capital P72U6 is warned ', due to last bit patterns and warn word difference too
Greatly, so confidence level is extremely low 0.0001.
And common seven in processing rule if the license plate of a vehicle is alert for capital P72U6, by presetting license plate type
The recognition result that the processing rule of position license plate obtains may be for ' (from empirically saying, model can be fuzzy ' police ' by capital P72U61 '
Word is considered as a vertical line, is judged as number 1), but since last position and ' 1 ' is not special picture, so confidence level is
0.3；And the obtained result of processing rule of the police car license plate in the processing rule by presetting license plate type is ' capital
P72U6 is alert ', and confidence level is 0.9.
Step S104 selects the hypothesis license plate corresponding to the maximum confidence level from multiple confidence levels
Recognition result is as target license plate recognition result.
It is set corresponding to the obtained hypothesis license plate recognition result of processing rule by comparing every kind of default license plate type
The size of reliability, to select the maximum hypothesis license plate recognition result of confidence level as target license plate recognition result.
Continue for example above-mentioned, be capital P72U6D for license plate, by the comparison of confidence level, selection is maximum
Hypothesis license plate recognition result corresponding to confidence level as final result, then final result will export result be ' capital P72U6D ',
I.e. target license plate recognition result is ' capital P72U6D '.And be capital P72U6 alert for license plate, since ' capital P72U6 is alert ' is corresponding
Confidence level is 0.9, and ' P72U6D ' corresponding confidence level in capital is 0.3, by selecting the maximum hypothesis Car license recognition knot of confidence level
Fruit is as final result, therefore final result will export ' capital P72U6 is alert '.
In order to more intuitively embody the beneficial effect of the licence plate recognition method in the embodiment of the present invention, as shown in Fig. 2, lead to
It crosses the processing rule by license plate recognition result Jing Guo every kind of default license plate type to handle, to obtain and every kind of default vehicle
The processing rule of board type is one-to-one to assume license plate recognition result, then by calculating each hypothesis license plate recognition result institute
Corresponding confidence level selects the maximum hypothesis license plate recognition result of confidence level as target license plate recognition result.So comparing
For the prior art, license plate recognition result is assumed correspondingly with the multiple default license plate type by generating, thus
So that xenogenesis license plate will retain higher confidence level, and confidence level will be big after through above-mentioned processing for the license plate of error type
It is big to reduce, and then by comparing the license plate recognition result of license plate after the processing rule for having passed through different default license plate types
Confidence level, and then effectively increase the accuracy to Car license recognition, to obtain the higher license plate recognition result of accuracy, gram
The technical problem low to Car license recognition accuracy in the presence of the prior art is taken.
As a kind of possible implementation, step S104 includes: the processing for default license plate type described in each
Rule determines that each hypothesis license plate recognition result medium size code bit includes the total confidence level of maximum of most two English alphabets；
Post-processing recognition result corresponding to the number position is determined according to maximum total confidence level；From with multiple default vehicles
It selects confidence level maximum in the corresponding multiple post-processing recognition results of processing rule of board type to know as target license plate
Other result.
Optionally, post-processing recognition result is different from license plate recognition result is assumed.
Optionally, the processing rule for each default license plate type determines each hypothesis vehicle
Board recognition result medium size code bit includes the total confidence level of maximum of most two English alphabets, comprising: described default for each
The processing rule of license plate type determines in number position to include first total confidence level of zero-bit English alphabet respectively, includes an English
Second total confidence level of text mother and the total confidence level of third comprising two English alphabets；By described first total confidence level, institute
It is maximum as maximum total confidence level to state numerical value in second total confidence level and the total confidence level of the third.
Optionally, the number position includes N characters, and the N is the integer more than or equal to 4, and described being directed to is each
The regular first total confidence level determined in number position comprising zero-bit English alphabet respectively of the processing of a default license plate type,
Second total confidence level comprising an English alphabet and the total confidence level of third comprising two English alphabets, comprising: for every
The processing rule of one default license plate type obtains confidence level most from confidence level corresponding to multiple preset numbers respectively
Big digital confidence level and the maximum alphabetical confidence level of acquisition confidence level from confidence level corresponding to multiple default letters；
First product of the first confidence level and the digital confidence level before obtaining in M comprising J English alphabets, the M are equal to institute
It states N and subtracts one, the J is the integer less than or equal to 2；Obtain the second confidence level in described preceding M comprising X English alphabets
With the second product of the alphabetical confidence level, the X is equal to the J and subtracts one；When J is 0, by first sum of products second
Numerical value is maximum in product is used as first total confidence level；It is when J is 1, numerical value in second product of the first sum of products is maximum
The total confidence level of conduct second；When J is 2, the maximum conduct third of numerical value in second product of the first sum of products is always set
Reliability.
Optionally, the first product meets: F [M] [j] * number confidence level, and the second product meets: F [M] [X] * letter is set
Reliability, wherein M=N-1, X=j-1, i.e. F [M] [j] are F [N-1] [j], and F [M] [X] is F [N-1] [j-1], the N
Indicate the length (or indicating that number position includes N characters) of number position.
Wherein, the first confidence level when including j English alphabets in N-1 number positions before F [M] [j] is indicated, F [N-
The second confidence level when including j-1 English alphabets in N-1 number positions before 1] [j-1] is indicated.
Then the first product and the second product maximum value meet: F [N] [j]=max (F [N-1] [j] * number confidence level, F
[N-1] [j-1] * letter confidence level)；
First total confidence level meets: F [N] [0]=max (F [N-1] [0] * number confidence level, F [N-1] [0-1] *
Alphabetical confidence level)；
Second total confidence level meets: F [N] [1]=max (F [N-1] [1] * number confidence level, F [N-1] [1-1] *
Alphabetical confidence level)；
The total confidence level of third meets: F [N] [2]=max (F [N-1] [2] * number confidence level, F [N-1] [2-1] *
Alphabetical confidence level).
In the present embodiment, by dynamic programming algorithm, the vehicle for enabling g [i] [j] to choose when representing selection F [i] [j] transfer
Board character shifts formula using following state:
F [i] [j]=max (confidence level of F [i-1] [j] * highest confidence level number, F [i-1] [j-1] * highest
The confidence level of confidence level letter)；
In F [L (number position total length)] [0], confidence level highest one is selected in F [L] [1], F [L] [2], passes through g
Final license plate recognition result can be obtained in [i] [j] reverse-direction derivation.For example, finding number position every by maximum total confidence level
Corresponding character, wherein character corresponding to number position every meets: g [N] [j]=f [N] [j]=max (F [N-1]
[j] * number confidence level, F [N-1] [j-1] * letter confidence level), for example, first character of number position meets: g [0]
[0]=f [0] [0]=max (F [0-1] [0] * number confidence level, F [0-1] [0-1] * letter confidence level), if g [0]
[0]=F [0-1] [0] * number confidence level, then it represents that first of number position is number, and the character of this is the number
Number corresponding to confidence level.If otherwise g [0] [0]=F [0-1] [0-1] * letter confidence level, then it represents that the of number position
One is letter, and the character of this is letter corresponding to the letter confidence level.And so on, to obtain number position institute
Some characters.
Optionally, the regular corresponding multiple post-processings of processing from multiple default license plate types
Select confidence level maximum as target license plate recognition result in recognition result, comprising: from multiple default license plate types
The corresponding multiple post-processing recognition results of processing rule in set corresponding to each post-processing recognition result of determination
Reliability；It selects the maximum post-processing recognition result of confidence level to identify as target license plate from multiple confidence levels to tie
Fruit.
For example, as shown in figure 3, by the number position of the hypothesis license plate recognition result that step S102 is obtained again into
Row processing (as by executing the possible implementation of one of step S104), respectively obtains and the multiple default license plate
The one-to-one recognition result of type determines confidence level corresponding to each recognition result；From multiple confidence levels
Select the maximum recognition result of confidence level as target license plate recognition result.By further to hypothesis Car license recognition knot
The number position of fruit is identified, it is possible to prevente effectively from unreasonable situation occurs in number position, as number position occurs three or more
The English alphabet of quantity further decreases the error rate of identification, effectively improves the recognition accuracy to license plate.
Licence plate recognition method provided by the embodiment of the present invention, by the way that license plate recognition result is passed through every kind of default license plate
The processing rule of type is handled, to obtain and the one-to-one hypothesis vehicle of the processing of every kind of default license plate type rule
Board recognition result, then by calculating confidence level corresponding to each hypothesis license plate recognition result, select the maximum hypothesis of confidence level
License plate recognition result is as target license plate recognition result.So compared to existing technologies, by generate with it is the multiple
Default license plate type assumes license plate recognition result correspondingly, so that xenogenesis license plate will retain higher confidence level,
And the license plate of error type confidence level after through above-mentioned processing will substantially reduce, and then by comparing the Car license recognition of license plate
As a result the confidence level after having passed through the processing rule of different default license plate types, and then effectively increase to Car license recognition
Accuracy, to obtain the higher license plate recognition result of accuracy, overcome in the presence of the prior art to Car license recognition standard
The low technical problem of exactness.
Second embodiment
Corresponding to the licence plate recognition method in first embodiment, Fig. 4 is shown to be known using license plate shown in first embodiment
The other one-to-one license plate recognition device of method.As shown in figure 4, the license plate recognition device 400 includes obtaining module 410, the
One data processing module 420, the second data processing module 430 and third data processing module 440.Wherein, obtain module 410,
The realization function and first of first data processing module 420, the second data processing module 430 and third data processing module 440
Corresponding step is poly- in embodiment corresponds, and to avoid repeating, the present embodiment is not described in detail one by one.
Module 410 is obtained, for obtaining the license plate recognition result identified by step-by-step recognition methods.
First data processing module 420, for according to multiple default license plate types processing rule and the Car license recognition
As a result determining to assume license plate recognition result correspondingly with the multiple default license plate type.
Optionally, the first data processing module 420 is also used to: determining the multiple default vehicle according to license plate recognition result
Board type processing rule institute correspondingly license plate hypothesis position；Determine the hypothesis position and the pervious institute in hypothesis position
There is the product of confidence level corresponding to the character of character bit；Determine character string corresponding to the maximum value in the product；According to
The text string generation and the multiple default license plate type assume license plate recognition result correspondingly.
Second data processing module 430, for determining the corresponding confidence of multiple hypothesis license plate recognition result institutes
Degree.
Third data processing module 440, for being selected corresponding to the maximum confidence level from multiple confidence levels
The hypothesis license plate recognition result as target license plate recognition result.
Optionally, third data processing module 440 is also used to: the processing for default license plate type described in each is advised
Then determine that each hypothesis license plate recognition result medium size code bit includes the total confidence level of maximum of most two English alphabets；Root
Post-processing recognition result corresponding to the number position is determined according to maximum total confidence level；From with multiple default license plates
It selects confidence level maximum in the corresponding multiple post-processing recognition results of processing rule of type to identify as target license plate
As a result.
Optionally, the determination target license plate recognition result medium size code bit includes most two English alphabets
Maximum total confidence level, comprising: determine first total confidence level in number position comprising zero-bit English alphabet respectively, include an English
Second total confidence level of text mother and the total confidence level of third comprising two English alphabets；By described first total confidence level, institute
It is maximum as maximum total confidence level to state numerical value in second total confidence level and the total confidence level of the third.
Optionally, the number position includes N characters, and the N is the integer more than or equal to 4, and described determines respectively
It include first total confidence level, second comprising an English alphabet total confidence level and the packet of zero-bit English alphabet in number position
The total confidence level of third containing two English alphabets, comprising: obtain confidence from confidence level corresponding to multiple preset numbers respectively
It spends maximum digital confidence level and obtains the maximum alphabetical confidence of confidence level from confidence level corresponding to multiple default letters
Degree；First product of the first confidence level and the digital confidence level before obtaining in M comprising J English alphabets, described M etc.
Subtract one in the N, the J is the integer less than or equal to 2；Obtain second setting comprising X English alphabets in described preceding M
Second product of reliability and the alphabetical confidence level, the X are equal to the J and subtract one；When J is 0, by first sum of products
Numerical value is maximum in second product is used as first total confidence level；When J is 1, by numerical value in second product of the first sum of products
It is maximum to be used as second total confidence level；When J is 2, the is used as by numerical value in second product of the first sum of products is maximum
Three total confidence levels.
Optionally, the regular corresponding multiple post-processings of processing from multiple default license plate types
Select confidence level maximum as target license plate recognition result in recognition result, comprising: from multiple default license plate types
The corresponding multiple post-processing recognition results of processing rule in set corresponding to each post-processing recognition result of determination
Reliability；Select the maximum recognition result of confidence level as target license plate recognition result from multiple confidence levels.
3rd embodiment
As shown in figure 5, being the schematic diagram of terminal device 300.The terminal device 300 includes memory 302, processor
304 and it is stored in the computer program 303 that can be run in the memory 302 and on the processor 304, the meter
The licence plate recognition method in first embodiment is realized when calculation machine program 303 is executed by processor 304, to avoid repeating, this
Place repeats no more.Alternatively, realizing Car license recognition described in second embodiment when the computer program 303 is executed by processor 304
The function of each model/unit in device, to avoid repeating, details are not described herein again.
Illustratively, computer program 303 can be divided into one or more module/units, one or more mould
Block/unit is stored in memory 302, and is executed by processor 304, to complete the present invention.One or more modules/mono-
Member can be the series of computation machine program instruction section that can complete specific function, and the instruction segment is for describing computer program
303 implementation procedure in terminal device 300.For example, computer program 303 can be divided into obtaining in second embodiment
Modulus block 410, the first data processing module 420, the second data processing module 430 and third data processing module 440, each mould
The concrete function of block will not repeat them here as described in the first embodiment or the second embodiment.
Terminal device 300 can be desktop PC, notebook, palm PC and cloud server etc. and calculate equipment.
Wherein, memory 302 may be, but not limited to, random access memory (Random Access Memory,
RAM), read-only memory (Read Only Memory, ROM), programmable read only memory (Programmable Read-
Only Memory, PROM), erasable read-only memory (Erasable Programmable Read-Only Memory,
EPROM), electricallyerasable ROM (EEROM) (Electric Erasable Programmable Read-Only Memory,
EEPROM) etc..Wherein, memory 302 is for storing program, and the processor 304 executes institute after receiving and executing instruction
Program is stated, the method for the flow definition that aforementioned any embodiment of the embodiment of the present invention discloses can be applied in processor 304,
Or it is realized by processor 304.
Processor 304 may be a kind of IC chip, the processing capacity with signal.Above-mentioned processor 304 can
To be general processor, including central processing unit (Central Processing Unit, CPU), network processing unit
(Network Processor, NP) etc.；Can also be digital signal processor (Digital Signal Processor,
DSP), specific integrated circuit (Application Specific Integrated Circuit, ASIC), ready-made programmable
Gate array (Field-Programmable Gate Array, FPGA) either other programmable logic device, discrete gate or
Transistor logic, discrete hardware components.It may be implemented or execute disclosed each method, the step in the embodiment of the present invention
Rapid and logic diagram.General processor can be microprocessor or the processor is also possible to any conventional processor etc..
It is understood that structure shown in fig. 5 is only a kind of structural schematic diagram of terminal device 300, terminal device
300 can also include than more or fewer components shown in Fig. 5.Each component shown in Fig. 5 can using hardware, software or
A combination thereof is realized.
Fourth embodiment
The embodiment of the present invention also provides a kind of storage medium, and instruction is stored on the storage medium, when described instruction exists
The Car license recognition side in first embodiment is realized when running on computer, when the computer program is executed by processor
Method, to avoid repeating, details are not described herein again.Alternatively, realizing second embodiment institute when the computer program is executed by processor
The function of each model/unit in license plate recognition device is stated, to avoid repeating, details are not described herein again.
Through the above description of the embodiments, those skilled in the art can be understood that the present invention can be with
By hardware realization, the mode of necessary general hardware platform can also be added to realize by software, based on this understanding,
Technical solution of the present invention can be embodied in the form of software products, which can store non-volatile at one
In property storage medium (can be CD-ROM, USB flash disk, mobile hard disk etc.), including some instructions are with so that a computer is set
The method that standby (can be personal computer, server or the network equipment etc.) executes each implement scene of the present invention.
The foregoing is only a preferred embodiment of the present invention, is not intended to restrict the invention, for the skill of this field
For art personnel, the invention may be variously modified and varied.All within the spirits and principles of the present invention, made any
Modification, equivalent replacement, improvement etc., should all be included in the protection scope of the present invention.It should also be noted that similar label and word
Mother indicates similar terms in following attached drawing, therefore, once it is defined in a certain Xiang Yi attached drawing, then in subsequent attached drawing
In do not need that it is further defined and explained.
Claims (10)
1. a kind of licence plate recognition method characterized by comprising
Obtain the license plate recognition result identified by step-by-step recognition methods；
According to the processing rule of multiple default license plate types and license plate recognition result determination and the multiple default license plate kind
Class assumes license plate recognition result correspondingly；
Determine the corresponding confidence level of multiple hypothesis license plate recognition result institutes；
Select the hypothesis license plate recognition result corresponding to the maximum confidence level as mesh from multiple confidence levels
Mark license plate recognition result.
2. the method according to claim 1, wherein the processing rule according to multiple default license plate types
License plate recognition result is assumed correspondingly with the multiple default license plate type with license plate recognition result determination, comprising:
According to license plate recognition result determine the multiple default license plate type processing rule institute correspondingly license plate hypothesis
Position；
Determine the product of confidence level corresponding to the hypothesis position and the character for assuming the pervious all character bits in position；
Determine character string corresponding to the maximum value in the product；
License plate recognition result is assumed correspondingly according to the text string generation and the multiple default license plate type.
3. the method according to claim 1, wherein described select maximum institute from multiple confidence levels
The hypothesis license plate recognition result corresponding to confidence level is stated as target license plate recognition result, comprising:
Each hypothesis license plate recognition result medium size code bit is determined for the processing rule of default license plate type described in each
The total confidence level of maximum comprising most two English alphabets；
Post-processing recognition result corresponding to the number position is determined according to maximum total confidence level；
Confidence is selected from the regular corresponding multiple post-processing recognition results of processing with multiple default license plate types
It spends maximum as target license plate recognition result.
4. according to the method described in claim 3, it is characterized in that, the place for each default license plate type
Reason rule determines that each hypothesis license plate recognition result medium size code bit includes the total confidence level of maximum of most two English alphabets,
Include:
For default license plate type described in each processing rule determine in number position respectively comprising zero-bit English alphabet the
One total confidence level, second comprising an English alphabet total confidence level and the total confidence level of third comprising two English alphabets；
Numerical value in described first total confidence level, second total confidence level and the total confidence level of the third is maximum as maximum
Total confidence level.
5. according to the method described in claim 4, it is characterized in that, the number position includes N characters, the N be greater than or
Integer equal to 4, the processing rule for each default license plate type determine to include zero in number position respectively
Position first total confidence level of English alphabet, second comprising an English alphabet total confidence level and comprising two English alphabets
The total confidence level of third, comprising:
It is obtained from confidence level corresponding to multiple preset numbers respectively for the processing rule of default license plate type described in each
It takes the maximum digital confidence level of confidence level and obtains the maximum word of confidence level from confidence level corresponding to multiple default letters
Female confidence level；
First product of the first confidence level and the digital confidence level before obtaining in M comprising J English alphabets, described M etc.
Subtract one in the N, the J is the integer less than or equal to 2；
The second product of the second confidence level and the alphabetical confidence level in described preceding M comprising X English alphabets is obtained, it is described
X is equal to the J and subtracts one；
It is when J is 0, numerical value in second product of the first sum of products is maximum as first total confidence level；
It is when J is 1, numerical value in second product of the first sum of products is maximum as second total confidence level；
It is when J is 2, numerical value in second product of the first sum of products is maximum as the total confidence level of third.
6. according to the method described in claim 3, it is characterized in that, the processing from multiple default license plate types
Select confidence level maximum as target license plate recognition result in the corresponding multiple post-processing recognition results of rule, comprising:
It is determined from the regular corresponding multiple post-processing recognition results of processing with multiple default license plate types each
Confidence level corresponding to the post-processing recognition result；
Select the maximum post-processing recognition result of confidence level as target license plate recognition result from multiple confidence levels.
7. a kind of license plate recognition device characterized by comprising
Module is obtained, for obtaining the license plate recognition result identified by step-by-step recognition methods；
First data processing module, for being determined according to the processing rule of multiple default license plate types and the license plate recognition result
License plate recognition result is assumed correspondingly with the multiple default license plate type；
Second data processing module, for determining the corresponding confidence level of multiple hypothesis license plate recognition result institutes；
Third data processing module, for selecting the vacation corresponding to the maximum confidence level from multiple confidence levels
License plate recognition result is determined as target license plate recognition result.
8. device according to claim 7, which is characterized in that the third data processing module is also used to:
Each hypothesis license plate recognition result medium size code bit is determined for the processing rule of default license plate type described in each
The total confidence level of maximum comprising most two English alphabets；
Post-processing recognition result corresponding to the number position is determined according to maximum total confidence level；
Confidence is selected from the regular corresponding multiple post-processing recognition results of processing with multiple default license plate types
It spends maximum as target license plate recognition result.
9. a kind of terminal device characterized by comprising memory, processor and storage are in the memory and can be
The computer program run on the processor, the processor realize such as claim 1 to 6 when executing the computer program
The step of described in any item licence plate recognition methods.
10. a kind of storage medium, which is characterized in that instruction is stored on the storage medium, when described instruction on computers
When operation, so that the computer executes such as licence plate recognition method as claimed in any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810899533.6A CN109063702B (en) | 2018-08-08 | 2018-08-08 | License plate recognition method, device, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810899533.6A CN109063702B (en) | 2018-08-08 | 2018-08-08 | License plate recognition method, device, equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109063702A true CN109063702A (en) | 2018-12-21 |
| CN109063702B CN109063702B (en) | 2022-05-17 |
Family
ID=64679032
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810899533.6A Active CN109063702B (en) | 2018-08-08 | 2018-08-08 | License plate recognition method, device, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109063702B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563504A (en) * | 2020-07-16 | 2020-08-21 | 平安国际智慧城市科技股份有限公司 | License plate recognition method and related equipment |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1085455B1 (en) * | 1999-09-15 | 2006-08-30 | Siemens Corporate Research, Inc. | License plate recognition with an intelligent camera |
| KR20070113334A (en) * | 2006-05-23 | 2007-11-29 | 주식회사 인프라밸리 | License plate recognition method using differential image and neural network |
| CN102693431A (en) * | 2012-05-31 | 2012-09-26 | 信帧电子技术（北京）有限公司 | Method and device for identifying type of white number plate |
| 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 |
| CN104915644A (en) * | 2015-05-29 | 2015-09-16 | 安徽四创电子股份有限公司 | License plate intelligent discrimination method based on difference data reconstruction |
| CN105631470A (en) * | 2015-12-21 | 2016-06-01 | 深圳市捷顺科技实业股份有限公司 | Method and system for verifying license plate type |
| CN106408950A (en) * | 2016-11-18 | 2017-02-15 | 北京停简单信息技术有限公司 | Parking lot entrance and exit license plate recognition system and method |
| CN106845478A (en) * | 2016-12-30 | 2017-06-13 | 同观科技（深圳）有限公司 | The secondary licence plate recognition method and device of a kind of character confidence level |
| US20170249524A1 (en) * | 2016-02-25 | 2017-08-31 | Xerox Corporation | Method and system for detection-based segmentation-free license plate recognition |
| US20170372161A1 (en) * | 2016-06-24 | 2017-12-28 | Accenture Global Solutions Limited | Intelligent automatic license plate recognition for electronic tolling environments |
| CN108108734A (en) * | 2016-11-24 | 2018-06-01 | 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 |
-
2018
- 2018-08-08 CN CN201810899533.6A patent/CN109063702B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1085455B1 (en) * | 1999-09-15 | 2006-08-30 | Siemens Corporate Research, Inc. | License plate recognition with an intelligent camera |
| KR20070113334A (en) * | 2006-05-23 | 2007-11-29 | 주식회사 인프라밸리 | License plate recognition method using differential image and neural network |
| CN102693431A (en) * | 2012-05-31 | 2012-09-26 | 信帧电子技术（北京）有限公司 | Method and device for identifying type of white number plate |
| 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 |
| CN104915644A (en) * | 2015-05-29 | 2015-09-16 | 安徽四创电子股份有限公司 | License plate intelligent discrimination method based on difference data reconstruction |
| CN105631470A (en) * | 2015-12-21 | 2016-06-01 | 深圳市捷顺科技实业股份有限公司 | Method and system for verifying license plate type |
| US20170249524A1 (en) * | 2016-02-25 | 2017-08-31 | Xerox Corporation | Method and system for detection-based segmentation-free license plate recognition |
| US20170372161A1 (en) * | 2016-06-24 | 2017-12-28 | Accenture Global Solutions Limited | Intelligent automatic license plate recognition for electronic tolling environments |
| CN106408950A (en) * | 2016-11-18 | 2017-02-15 | 北京停简单信息技术有限公司 | Parking lot entrance and exit license plate recognition system and method |
| CN108108734A (en) * | 2016-11-24 | 2018-06-01 | 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 |
| CN106845478A (en) * | 2016-12-30 | 2017-06-13 | 同观科技（深圳）有限公司 | The secondary licence plate recognition method and device of a kind of character confidence level |
Non-Patent Citations (4)
| Title |
|---|
| DIVYA GILLY 等: "License Plate Recognition- A Template Matching Method", 《INTERNATIONAL JOURNAL OF ENGINEERING RESEARCH AND APPLICATIONS (IJERA)》 * |
| G. ANGEL 等: "ALPR for Ambiguous Character using Template Matching with Fuzzy Classifiers: a Survey", 《INTERNATIONAL JOURNAL OF COMPUTING ACADEMIC RESEARCH (IJCAR)》 * |
| 何铁军 等: "车牌识别算法的研究与实现", 《公路交通科技》 * |
| 顾弘 等: "车牌识别中先验知识的嵌入及字符分割方法", 《中国图象图形学报》 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111563504A (en) * | 2020-07-16 | 2020-08-21 | 平安国际智慧城市科技股份有限公司 | License plate recognition method and related equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109063702B (en) | 2022-05-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107682348A (en) | DGA domain name Quick method and devices based on machine learning | |
| CN111160140B (en) | Image detection method and device | |
| WO2022077907A1 (en) | Adversarial attack detection method, system and device, and computer-readable storage medium | |
| CN111753290A (en) | Software type detection method and related equipment | |
| CN113961438A (en) | Multi-granularity and multi-hierarchy based historical behavior abnormal user detection system, method, equipment and storage medium | |
| CN116167010A (en) | Rapid identification method for abnormal events of power system with intelligent transfer learning capability | |
| CN109670525A (en) | Object detection method and system based on once shot detection | |
| CN109063702A (en) | Licence plate recognition method, device, equipment and storage medium | |
| CN112037174B (en) | Chromosome abnormality detection method, chromosome abnormality detection device, chromosome abnormality detection apparatus, and computer-readable storage medium | |
| CN111340139B (en) | Method and device for judging complexity of image content | |
| CN111382052A (en) | Code quality evaluation method and device and electronic equipment | |
| CN113705201B (en) | Text-based event probability prediction evaluation algorithm, electronic device and storage medium | |
| CN114925765A (en) | Construction method, device, equipment and storage medium of antagonism integrated classification model | |
| CN115393617A (en) | Simulated trademark rapid detection method and system based on multi-convolution kernel inspection | |
| CN115189961A (en) | Fault identification method, device, equipment and storage medium | |
| CN114398887A (en) | Text classification method and device and electronic equipment | |
| CN111027429A (en) | Data preprocessing method and system for intelligent identification of electrical drawings | |
| CN113111833B (en) | Safety detection method and device of artificial intelligence system and terminal equipment | |
| KR102588409B1 (en) | Vehicle license plate detection device, method and program | |
| CN115098566B (en) | Information system for improving convolutional neural network model | |
| CN113743110B (en) | Word missing detection method and system based on fine-tuning generation type confrontation network model | |
| CN117807481B (en) | Fault identification method, training device, training equipment and training medium | |
| CN110163471B (en) | Abnormality identification method and device | |
| CN109376739B (en) | Marshalling mode determining method and device | |
| CN117454147A (en) | Training method and device for anomaly detection model |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |