CN109492642A - Licence plate recognition method, device, computer equipment and storage medium - Google Patents
Licence plate recognition method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109492642A CN109492642A CN201811113599.4A CN201811113599A CN109492642A CN 109492642 A CN109492642 A CN 109492642A CN 201811113599 A CN201811113599 A CN 201811113599A CN 109492642 A CN109492642 A CN 109492642A
- Authority
- CN
- China
- Prior art keywords
- character
- license plate
- image
- target
- sliding window
- 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
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000003860 storage Methods 0.000 title claims abstract description 17
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 14
- 238000003708 edge detection Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 18
- 238000004422 calculation algorithm Methods 0.000 claims description 17
- 238000005520 cutting process Methods 0.000 claims description 12
- 238000007781 pre-processing Methods 0.000 claims description 10
- 230000011218 segmentation Effects 0.000 claims description 9
- 238000000638 solvent extraction Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 8
- 229910052704 radon Inorganic materials 0.000 claims description 8
- SYUHGPGVQRZVTB-UHFFFAOYSA-N radon atom Chemical compound [Rn] SYUHGPGVQRZVTB-UHFFFAOYSA-N 0.000 claims description 8
- 238000012937 correction Methods 0.000 claims description 7
- 230000005484 gravity Effects 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 6
- 230000005764 inhibitory process Effects 0.000 claims description 6
- 230000009977 dual effect Effects 0.000 claims description 5
- 235000013399 edible fruits Nutrition 0.000 claims description 3
- 230000006870 function Effects 0.000 description 11
- 230000008859 change Effects 0.000 description 5
- 238000010606 normalization Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 238000012544 monitoring process Methods 0.000 description 4
- 230000005540 biological transmission Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 210000004027 cell Anatomy 0.000 description 2
- 238000005260 corrosion Methods 0.000 description 2
- 230000007797 corrosion Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 235000019082 Osmanthus Nutrition 0.000 description 1
- 241000333181 Osmanthus Species 0.000 description 1
- 230000033228 biological regulation Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 235000009508 confectionery Nutrition 0.000 description 1
- 230000010485 coping Effects 0.000 description 1
- 238000009795 derivation Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000010977 jade Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 210000003296 saliva Anatomy 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Character Discrimination (AREA)
- Character Input (AREA)
Abstract
The invention discloses a kind of licence plate recognition methods, device, computer equipment and storage medium, accompanying method includes: by obtaining initial license plate image, and initial license plate image is pre-processed, obtain target license plate image, and then use preset sliding window, carry out slide from left to right on target license plate image, sliding every time, the target license plate image within the scope of sliding window is obtained as a sliding window image, so that it includes the sliding window image that partly overlaps that license plate image, which is divided into multiple, character in license plate number occurs repeatedly, avoid the unrecognized situation of partial character in license plate number caused by manually dividing to target license plate image, improve the versatility of Car license recognition, obtained sliding window image is input in convolutional neural networks and is identified, obtain recognition result, in adjacent identification As a result there are multiple duplicate characters in, be conducive to correctly extract the character in license plate number, improve the accuracy rate of Car license recognition.
Description
Technical field
The present invention relates to technical field of image processing fields more particularly to a kind of licence plate recognition method, device, computer to set
Standby and storage medium.
Background technique
With the development of the social economy, more and more automobiles appear in road traffic or parking facility, it gives people
Life bring many conveniences, but the management of automobile also becomes to become increasingly complex.Such as vehicle toll and management, the magnitude of traffic flow
Detection, parking lot fee collection management, monitoring vehicle breaking regulation, the particular problems such as fake license vehicle identification.
For these problems, currently employed main method is to be managed by identification license plate to vehicle, current main
The license plate recognition technology to be used be license plate image is split by direct artificial delimitation ratio, and then to segmentation after
Image carries out Car license recognition, and this method recognition accuracy is low, and can only identify the license plate of prespecified size, in shooting
When the angle of license plate image is unstable or not of uniform size, unrecognized situation can be generated, does not have robustness and general
Property.
Summary of the invention
The embodiment of the present invention provides a kind of licence plate recognition method, device, computer equipment and storage medium, current to solve
Car license recognition accuracy rate is low and the weak problem of versatility.
A kind of licence plate recognition method, comprising:
Obtain initial license plate image；
The initial license plate image is pre-processed, the target license plate image that size is a × b pixel is obtained, wherein
A, b is positive integer；
Use default size for the sliding window of c × b pixel, using 1 pixel as step-length, in the target license plate image
On carry out slide from left to right, every time slide, the target license plate image within the scope of the sliding window is obtained, as one
Sliding window image, wherein c < a and c are positive integer；
After sliding a-c times, slide is terminated, a-c sliding window image is obtained；
The a-c sliding window images are input in convolutional neural networks and are identified, a-c identification knot is obtained
Fruit, wherein each recognition result includes several characters；
All characters in the a-c recognition results are formed into character string, according to default partitioning scheme to the character
String is split, and target character is obtained from obtained each substring, and determine target license plate according to the target character
Number.
A kind of license plate recognition device, comprising:
Module is obtained, for obtaining initial license plate image；
Preprocessing module obtains the mesh that size is a × b pixel for pre-processing to the initial license plate image
Mark license plate image, wherein a, b are positive integer；
Sliding block, for using default size for the sliding window of c × b pixel, using 1 pixel as step-length, in institute
It states and carries out slide on target license plate image from left to right, slide every time, obtain the target carriage within the scope of the sliding window
Board image, as a sliding window image, wherein c < a and c are positive integer；
Module is terminated, after sliding a-c times, slide is terminated, obtains a-c sliding window image；
Identification module is identified for the a-c sliding window images to be input in convolutional neural networks, is obtained
A-c recognition result, wherein each recognition result includes several characters；
Divide module, for all characters in the a-c recognition results to be formed character string, according to default segmentation side
Formula is split the character string, target character is obtained from obtained each substring, and according to the target character
Determine target license plate number.
A kind of computer equipment, including memory, processor and storage are in the memory and can be in the processing
The computer program run on device, the processor realize the step of above-mentioned licence plate recognition method when executing the computer program
Suddenly.
A kind of computer readable storage medium, the computer-readable recording medium storage have computer program, the meter
The step of calculation machine program realizes above-mentioned licence plate recognition method when being executed by processor.
Above-mentioned licence plate recognition method, device, computer equipment and storage medium, on the one hand, by obtaining initial license plate figure
Picture, and initial license plate image is pre-processed, the target license plate image that size is a × b pixel is obtained, and then using default
Size is that the sliding window of c × b pixel carries out sliding behaviour using 1 pixel as step-length from left to right on target license plate image
Make, slide every time, obtain the target license plate image within the scope of sliding window, as a sliding window image, in sliding a-c times
Afterwards, a-c sliding window image is obtained, slide is terminated, so that it includes the sliding that partly overlaps that license plate image, which is divided into multiple,
Video in window, the character in license plate number occur repeatedly, avoid relying on license plate caused by the Character segmentation template of artificial division
The unrecognized situation of partial character in number, improves the versatility of Car license recognition；On the other hand, by obtained a-c
Sliding window image, which is input in convolutional neural networks, to be identified, a-c recognition result is obtained, and is deposited in adjacent recognition result
In multiple duplicate characters, be conducive to correctly extract the character in license plate number, meanwhile, the character from a-c recognition result
In extract target character, and then determine target license plate number, improve the accuracy rate of Car license recognition.
Detailed description of the invention
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below by institute in the description to the embodiment of the present invention
Attached drawing to be used is needed to be briefly described, it should be apparent that, the accompanying drawings in the following description is only some implementations of the invention
Example, for those of ordinary skill in the art, without any creative labor, can also be according to these attached drawings
Obtain other attached drawings.
Fig. 1 is the application environment schematic diagram of licence plate recognition method provided in an embodiment of the present invention；
Fig. 2 is the implementation flow chart of licence plate recognition method provided in an embodiment of the present invention；
Fig. 3 is the implementation flow chart of step S20 in licence plate recognition method provided in an embodiment of the present invention；
Fig. 4 is the implementation flow chart of step S21 in licence plate recognition method provided in an embodiment of the present invention；
Fig. 5 is the implementation flow chart of step S60 in licence plate recognition method provided in an embodiment of the present invention；
Fig. 6 is the implementation flow chart of step S64 in licence plate recognition method provided in an embodiment of the present invention；
Fig. 7 is the schematic diagram of license plate recognition device provided in an embodiment of the present invention；
Fig. 8 is the schematic diagram of computer equipment provided in an embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are some of the embodiments of the present invention, instead of all the embodiments.Based on this hair
Embodiment in bright, every other implementation obtained by those of ordinary skill in the art without making creative efforts
Example, shall fall within the protection scope of the present invention.
Referring to Fig. 1, Fig. 1 shows the application environment of licence plate recognition method provided in an embodiment of the present invention.The Car license recognition
Method is applied to be carried out in Car license recognition scene in the license plate for the vehicle taken.The identification scene includes server-side and client
End, wherein be attached between server-side and client by network, client will take license plate image and be sent to service
End, server-side receive the license plate image that client is sent and are identified, client specifically can be, but not limited to be hypervelocity camera shooting
It head, the monitoring of day net, electronic police, various personal computers, laptop, smart phone, tablet computer and portable wears
Equipment is worn, server-side can specifically be realized with the server cluster that independent server or multiple servers form.
Referring to Fig. 2, being applied in this way Fig. 2 shows a kind of licence plate recognition method provided in an embodiment of the present invention in Fig. 1
In server-side for be illustrated, details are as follows:
S10: initial license plate image is obtained.
Specifically, after client obtains initial license plate image, initial license plate image is sent to by the network transmission protocol
Server-side, server-side receive the initial license plate image by the network transmission protocol.
Wherein, client specifically can be the monitoring with shooting function such as hypervelocity camera, the monitoring of day net, electronic police
Equipment, directly shooting obtain initial license plate image, are also possible to various personal computers, laptop, smart phone or flat
The intelligent terminals such as plate computer, the memory space with storing initial license plate image, and network interaction is carried out with server-side
Function.
Wherein, the network transmission protocol includes but is not limited to: Internet Control Message agreement (Internet Control
Message Protocol, ICMP), address resolution protocol (ARP Address Resolution Protocol, ARP) and text
Part transport protocol (File Transfer Protocol, FTP) etc..
S20: pre-processing initial license plate image, obtains the target license plate image that size is a × b pixel, wherein
A, b is positive integer.
Specifically, due to the angle of shooting, distance and automobile the factors such as run at high speed influence, get just
Beginning license plate image is of low quality, and directly progress recognition correct rate is lower, it is then desired to first be located initial license plate image in advance
Reason influences to reduce these factor brings, improves the accuracy rate of subsequent identification, pre-process to initial license plate image
Afterwards, the target license plate image that size is a × b pixel is obtained.
Wherein, pretreatment includes but is not limited to: image cropping, normalization and slant correction etc..
Wherein, a and b is preset numerical value, and basic unit is the width of 1 pixel, using a as horizontal direction side length, i.e.,
Using the width of a pixel as horizontal direction side length, using b as vertical direction side length, i.e., using the width of b pixel as vertical side
To side length, a and b are positive integer, and specific value, which can according to need, to be configured, and is not specifically limited herein.
Preferably, in embodiments of the present invention, the value of a takes the value of 140, b to take 28, that is, finally obtaining pixel is 140 × 28
Target license plate image.
S30: use default size for the sliding window of c × b pixel, using 1 pixel as step-length, in target license plate image
On carry out slide from left to right, every time slide, obtain sliding window within the scope of a target license plate image, as one
Sliding window image, wherein c < a and c are positive integer.
Specifically, by presetting the sliding window that size is c × b pixel, using 1 pixel as step-length, in target license plate
Slide is carried out on image from left to right, is slided every time, the target license plate image within the scope of sliding window is obtained, as one
Sliding window image so that multiple sliding windows have the image section of phase negative lap after multiple sliding, be conducive to it is subsequent into
Row target character it is preferred.
Wherein, the width b of sliding window and target license plate image is of same size, length c can according to the actual situation into
Row setting, is not particularly limited herein.
Preferably, the numerical value of c is set as 28 in the embodiment of the present invention, i.e., is made using the square of 28 × 28 pixels
It is slided for sliding window.
S40: after sliding a-c times, slide is terminated, a-c sliding window image is obtained.
Specifically, the length of target license plate image is a, and the length of sliding window is b, and sliding step 1 is passing through a-c
The right end of secondary sliding, sliding window is Chong Die with the right end of target license plate image, i.e., sliding window moves target license plate image most
At this time one a-c sliding window image is obtained in right end.
S50: a-c sliding window image being input in convolutional neural networks and is identified, obtains a-c identification knot
Fruit, wherein each recognition result includes several characters.
Specifically, a-c sliding window image obtained in step S40 is input to trained as input picture
It is identified in convolutional neural networks, in the full articulamentum of convolutional neural networks, is preset with 71 classifiers, each classifier pair
Answer a preset characters, totally 71 preset characters, be respectively as follows: preset 35 Chinese characters, preset 26 capitalization English letters and
Preset 10 Arabic numerals, sliding window image are input to full articulamentum after convolution algorithm, use preset 71
Classifier carries out Classification and Identification, obtains recognition result.
It is worth noting that each recognition result is the character or multiple characters in this 71 preset characters.
Wherein, preset 35 Chinese characters include: capital, saliva, Ji, Shanxi, illiteracy, the Liao Dynasty, Ji, black, Shanghai, Soviet Union, Zhejiang, Anhui, Fujian, Jiangxi, Shandong,
Henan, Hubei Province, Hunan, Guangdong, osmanthus, fine jade, Chongqing, river, Guizhou Province, Yunnan, hiding, Shan, sweet, green, hiding, peaceful, new, platform, port, Australia.
Wherein, preset 26 capitalization English letters include: A, B, C, D, E, F, G, H, I, J, K, L, M, N, O, P, Q, R,
S、T、U、V、W、X、Y、Z。
Wherein, preset 10 Arabic numerals include: 0,1,2,3,4,5,6,7,8,9.
Wherein, convolutional Neural net (Convolutional Neural Network, CNN) is a kind of feedforward neural network,
Its artificial neuron can respond the surrounding cells in a part of coverage area, can rapidly and efficiently carry out image procossing.
S60: by a-c recognition result all characters form character string, according to default partitioning scheme to character string into
Row segmentation, obtains target character, and determine target license plate number according to target character from obtained each substring.
Specifically, all characters in a-c recognition result are formed character string, and according to default partitioning scheme to word
Symbol string is split, and obtains at least two substrings, and then preferably goes out a target character from each substring, is passed through
Each target character is combined, target license plate number is obtained.
Wherein, default partitioning scheme can be the substring quantity divided as needed, carry out average mark to character string
It cuts, is also possible to be split substring according to preset range or default cut-point, can also carry out according to the actual situation
Setting, is not especially limited herein.
It is worth noting that since license plate number is generally 7 characters, the embodiment of the present invention as a preferred method,
The character string that a-c recognition result forms is divided into 7 substrings according to default partitioning scheme.
In the present embodiment, by obtaining initial license plate image, and initial license plate image is pre-processed, obtaining size is
The target license plate image of a × b pixel, and then use default size for the sliding window of c × b pixel, it is step with 1 pixel
It is long, it carries out slide from left to right on target license plate image, slides every time, obtain the target within the scope of the sliding window
License plate image after sliding a-c times, obtains a-c sliding window image as a sliding window image, terminates sliding behaviour
Make, so that it includes the sliding window image that partly overlaps that license plate image, which is divided into multiple, the character in license plate number occurs repeatedly, keeps away
The unrecognized situation of partial character in license plate number caused by relying on the Character segmentation template of artificial division is exempted from, has improved
The versatility of Car license recognition, and then a-c obtained sliding window image is input in convolutional neural networks and is identified, it obtains
To a-c recognition result, there are multiple duplicate characters in adjacent recognition result, be conducive to correctly extract in license plate number
Character, meanwhile, target character is extracted from the character in the a-c recognition results, and then determine target license plate number, mention
The high accuracy rate of Car license recognition.
On the basis of the corresponding embodiment of Fig. 2, below by a specific embodiment come to being mentioned in step S20
And initial license plate image is pre-processed, obtain size be a × b pixel target license plate image concrete methods of realizing
It is described in detail.
Referring to Fig. 3, Fig. 3 shows the specific implementation flow of step S20 provided in an embodiment of the present invention, details are as follows:
S21: by edge detection algorithm, the lower boundary of the coboundary of license plate and license plate in initial license plate image is obtained.
Specifically, since the factors such as the angle and distance of shooting influence, the license plate image taken is generally in addition to license plate area
Except domain, it will also include some non-license plate images outside license plate area, these non-license plate images can make subsequent Car license recognition
It thus in order to improve the accuracy rate of subsequent Car license recognition, needs to find out initial license plate figure by edge inspection algorithms at interference
The coboundary of license plate and the lower boundary of license plate as in, so that it is determined that the license plate range in initial license plate image.
Wherein, edge detection algorithm is the algorithm for carrying out edge detection, and edge detection is image procossing and computer
Basic problem in vision, in embodiments of the present invention, the purpose of edge detection are that brightness change is obvious in mark license plate image
Point, i.e. the point on license plate boundary, the significant changes in image attributes usually reflect the critical event and variation of attribute, these figures
As attribute includes but is not limited to: discontinuous, surface direction in depth is discontinuous, material property variation and scene lighting variation
Deng.
Wherein, the edge of image refers to that region jumpy occurs for gray scale in image, and the situation of change of ganmma controller can
To be reflected with the gradient of intensity profile.
Common edge detection algorithm includes but is not limited to: Sobel Operator (Sobel operator) edge detection is calculated
Method, Gauss-Laplace edge detection algorithm, Luo Baici overlapping edges detection (Roberts Cross operator) are calculated
Method and Canny multistage edge detection algorithm etc..
Preferably, the edge detection algorithm used by inventive embodiments is Canny multistage edge detection algorithm.
It is worth noting that the coboundary of obtained license plate and the lower boundary of license plate are two sections according to edge detection algorithm
Line segment.That is, the coboundary for obtaining license plate includes two vertex in left and right, the lower boundary of obtained license plate includes two vertex in left and right.
S22: according to the lower boundary of the coboundary of license plate and license plate, the range image of license plate is determined.
Specifically, the lower boundary of the coboundary of license plate obtained in step S21 and license plate is connected, that is, by license plate
The left side vertex of coboundary be connected with the left side vertex of the lower boundary of license plate, by the right vertex of the coboundary of license plate and license plate
Lower boundary the right vertex be connected, so that a quadrangle is obtained, using the image in this square range as license plate
Range image.
S23: slant correction is carried out to range image using Radon transform, the base image after being corrected.
Specifically, due to the influence of the angle and distance of shooting, the initial license plate image got can have inclination, in turn
There is also inclinations for obtained range image, in order to improve the accuracy rate of subsequent Car license recognition, by Radon transform to range image
Slant correction is carried out, thus the base image after being corrected.
Wherein, Radon transform (Radon transform) is that one kind is superimposed by determining direction projection, finds maximal projection value
When angle, so that it is determined that image inclination angle, and then be corrected, the method for the image after being corrected.
S24: centered on the center of gravity of base image, cutting base image, and obtaining size is a × b pixel
Target license plate image.
Specifically, right using b as vertical direction side length using a as horizontal direction side length centered on the center of gravity of base image
The base image that step S23 is obtained is cut, and the target license plate image of the rectangle of an a × b pixel size is obtained.
Preferably, the preset a of the embodiment of the present invention is 140, b 28, that is, the target of 140 × 28 pixels finally obtained
License plate image.
For example, in a specific embodiment, the coordinate of the center of gravity of base image is (82,21), and horizontal direction side length is
140, vertical direction side length is 28, then obtains top left corner apex coordinate (12,7), upper right corner apex coordinate (152,7), the lower left corner
Apex coordinate (12,35), lower right corner apex coordinate (152,35), by the rectangle for 140 × 28 pixels that this four vertex form,
Along this clipping rectangle base image, the image in the rectangular extent, i.e. target license plate image are obtained.
Preferably, after obtaining target license plate image, the embodiment of the present invention also carries out mean value to target license plate image
And normalization, eliminate the difference in target license plate image between different dimensions image data.
Wherein, normalization refers to the characteristics of image amplitude normalization in target license plate image to same range, even if
The standard deviation that all characteristics of image are removed with each characteristics of image, using obtained result as the image after characteristics of image normalization
Feature.
Wherein, it goes mean value to refer to and each dimension of characteristics of image all centers in target license plate image is turned to 0, i.e., by target
The central point of license plate image is withdrawn on coordinate origin.
In the present embodiment, by edge detection algorithm, obtain in initial license plate image under the coboundary and license plate of license plate
Boundary, and then according to the lower boundary of the coboundary of license plate and license plate, it determines the range image of license plate, reuses Radon transform to model
It encloses image and carries out slant correction, the base image after being corrected, then centered on the center of gravity of base image, to base image
It is cut, obtains the target license plate image that size is a × b pixel, when so that subsequent use base image being identified, kept away
Exempt from identification error caused by interference and the Tilt factor as the image outside license plate range, enhances the matter of target license plate image
Amount, is conducive to the accuracy rate for improving subsequent Car license recognition.
On the basis of the corresponding embodiment of Fig. 3, below by a specific embodiment come to being mentioned in step S21
And by edge detection algorithm, obtain the specific implementation side of the lower boundary of the coboundary and license plate of license plate in initial license plate image
Method is described in detail.
Referring to Fig. 4, Fig. 4 shows the specific implementation flow of step S21 provided in an embodiment of the present invention, details are as follows:
S211: noise removal is carried out to initial license plate image by Gaussian Blur, obtains denoising license plate image.
Specifically, the edge of license plate image is high-frequency signal, but the noise of image also focuses on high-frequency signal, it is easy to quilt
It is mistakenly identified as edge, it is then desired to remove the noise of image, avoids the noise of image to determining edge bring interference.At this
In inventive embodiments, noise removal is carried out to initial license plate image using Gaussian Blur, obtains denoising license plate image.
Wherein, the edge of license plate image refers to that gray scale occurs sharply in the image of license plate area and non-license plate area intersection
The region of variation.The situation of change of ganmma controller can be reflected with the gradient of intensity profile.
Wherein, the noise of the noise of image namely image, refer to be present in it is unnecessary or extra in image data
Interference information, the presence of noise have seriously affected the quality of image, therefore before image enhancement processing and classification processing, it is necessary to
Corrected, in embodiments of the present invention, the correcting method used for use Gaussian Blur remove noise.
Wherein, Gaussian Blur (Gaussian Blur), is also Gaussian smoothing, it is a kind of image fuzzy filter, it is used
Normal distribution calculates the transformation of each pixel in image, and picture noise is usually reduced with it and reduces level of detail.
It is worth noting that image border and noise are high-frequency signal, thus the radius selection of Gaussian Blur is critically important,
Excessive radius is easy to allow some weak endpoint detections less than the specific setting of radius can be adjusted according to the actual situation
It is whole, herein with no restriction.
S212: the gradient horizontally and vertically of denoising license plate image is calculated using preset gradient operator
Value, obtains Initial Gradient value set.
Specifically, the edge of image can be pointed in different directions, and by using preset operator, calculate denoising license plate image
Gradient value horizontally and vertically, obtain Initial Gradient value set.
Wherein, digital picture is exactly discrete point value spectrum, can also be two-dimensional discrete function, the gradient of image is exactly this
The derivation of two-dimensional discrete function as a result, gradient operator, that is, the method for being used to calculate gradient.
Wherein, preset gradient operator includes but is not limited to: Sobel Operator (Sobel operator), Prewitt are calculated
Son, Luo Baici operator (Roberts operator) and Canny operator.
Preferably, gradient operator used in the embodiment of the present invention is Canny operator.
S213: edge thinning processing is carried out to Initial Gradient value set by the way of the inhibition of non-maximum value, obtains one
The wide gradient edge of pixel.
Specifically, edge thinning processing is carried out to Initial Gradient value set by the way of the inhibition of non-maximum value, obtains one
The wide gradient edge of a pixel.
Wherein, non-maximum value inhibit (Non Maximum Suppression, NMS) be inhibit be not maximum element,
It can be regarded as part and carry out maximum value search, to help to retain local maxima gradient and inhibit every other gradient value, this meaning
Taste only remain position most sharp keen in change of gradient.
For example, in a specific embodiment, in vertical direction, thering is the gradient value of 4 pixel wides to constitute an office
Portion searches out the maximum picture of gradient value in the gradient value of this part in lid part by the way of the inhibition of non-maximum value
Vegetarian refreshments, as gradient edge, to realize edge thinning.
S214: using the weak marginal point in preset dual threshold filter gradient edge, the strong edge in gradient edge is obtained
Point.
Specifically, edge pixel is distinguished by the way that dual threshold, i.e. a high threshold values and a low valve valve is arranged.If edge
Pixel gradient value is greater than high threshold values, then is considered as strong edge point, if edge gradient value is less than high threshold values, is greater than low valve
Value is then labeled as weak marginal point, and the point less than low valve valve is then suppressed, and strong edge point can be confirmed to be genuine edge, but
Weak marginal point then may be genuine edge, it is also possible to accurate as a result, in this hair to obtain caused by noise or color change
In bright embodiment, weak marginal point is also suppressed.
S215: according to strong edge point, the coboundary of license plate and the lower boundary of license plate are determined.
Specifically, it is obtained under the coboundary and license plate of license plate according to strong edge point by image procossings such as corrosion extensions
Boundary.
Wherein, corrosion can eliminate independent point in strong edge point, and extension can be by adjacent but disjunct strong edge point
It connects.
In the present embodiment, noise removal is carried out to initial license plate image by Gaussian Blur, obtains denoising license plate image, and
The gradient value horizontally and vertically that denoising license plate image is calculated using preset gradient operator, obtains Initial Gradient
Value set, and then edge thinning is carried out to Initial Gradient value set by the way of the inhibition of non-maximum value, it is wide to obtain a pixel
Gradient edge obtain the strong edge point in gradient edge using the weak marginal point in preset dual threshold filter gradient edge,
Then according to strong edge point, the coboundary of license plate and the lower boundary of license plate are determined, the accuracy rate of license plate edge detection is improved, has
Conducive to the subsequent determination to license plate range.
On the basis of the corresponding embodiment of Fig. 2, below by a specific embodiment come to being mentioned in step S60
And by a-c recognition result all characters form character string, character string is split according to default partitioning scheme, from
Target character is obtained in obtained each substring, and the concrete methods of realizing of target license plate number is determined according to target character
It is described in detail.
Referring to Fig. 5, Fig. 5 shows the specific implementation flow of step S60 provided in an embodiment of the present invention, details are as follows:
Character in a-c recognition result: being successively put into character set X by S61, obtains X={ x1, x2..., xq,
In, q is the sum for the character for including, x in a-c recognition resultiFor i-th of character in character set X, i ∈ [1, q], q is positive
Integer.
Specifically, the character for including in each recognition result is sequentially placed into initial value is to obtain X in empty character set X
={ x1, x2..., xq, wherein q is the sum for the character for including, x in a-c recognition resultiFor i-th in character set X
Character.
For example, in a specific embodiment, having obtained 112 recognition results, wherein preceding 20 recognition results are as follows: " ",
" I ", " L ", " L ", " O ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai, I ", " B ", " B ",
" B ", " B ", " B ", wherein the content in each double quotation marks is a recognition result, and each recognition result can be 1 character,
Or sky, or multiple characters, this 20 recognition results are sequentially stored into character set X obtain X=" I ",
" L ", " L ", " O ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " I ", " B ", " B ",
“B”、“B”、“B”}。
S62: according to k+1 preset range, the k cut-off of character set X is determined, wherein k < q, and k is positive integer.
Specifically, it after step S61 obtains character set X, needs character set X carrying out cutting according to target character number
A target character is determined at multiple sub- character set, and then according to every sub- character set.
In embodiments of the present invention, it is assumed that according to the character quantity k+1 of license plate to be identified, be previously provided with k+1 model
It encloses, then determines k cut-off in character set X according to k+1 preset range.
Wherein, preset range is used to limit the preferred scope of each target character, for example, first preset range be (0,
15], i.e., using first recognition result to the 15th recognition result as first preset range.
By taking the example in step S61 as an example, if the first preset range be (0,15], obtained character set X=" I ", " L ",
" L ", " O ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " Shanghai ", " I ", " B ", " B ", " B ",
" B ", " B " }, then it chooses the 15th recognition result and is used as cut-off later, i.e., by the 15th element " I " of character set X and the 16th
Position between a element " B " is as first cut-off.
It is worth noting that preset range may include decimal, for example, second preset range be (15,31.5], i.e., the
Before two preset ranges include the 16th recognition result to all characters and second recognition result of the 31st recognition result
50% character.
Wherein, target character refers to the character in license plate number character, and current most of license plate number is 7 target words
Symbol perhaps the number of 8 target characters thus target character be usually 7 or 8, also can be set according to actual needs for
Other numerical value, are not specifically limited herein.
S63: character cutting is carried out to q character in character set X using k cut-off, obtains k+1 sub- character set.
Specifically, character cutting is carried out to the character in character set X using k obtained cut-off, obtains k+1 sub- words
Symbol collects to arrive k+1 preferred scope.
Wherein, cutting refers to using two adjacent cut-offs as boundary, by all characters in two adjacent cutting point ranges
As a sub- character set, first cut-off is made first character collection to all characters in the cutting point range
For first sub- character set, for the last one cut-off, by all of the last one cut-off to a last character range
Character is as last sub- character set.
S64: the most character of frequency of occurrence obtains k+1 target word as target character in every sub- character set of acquisition
Symbol.
Specifically, for every sub- character set, the number occurred to character each in the sub- character set is counted, and will be gone out
Target character of the most character of occurrence number as the sub- character set, obtains k+1 target character.
For example, in a specific embodiment, one of them sub- character set be " B ", " B ", " B ", " B ", " L ", " B ",
" B ", " B ", " B ", " O ", " L ", " B ", " B ", " B ", " B " }, it is counted, obtaining result is comprising 12 characters " B ", two words
Accord with " L " and a character " O ", the target character by character " B " as the sub- character set.
S65: according to the sequence of k+1 preset range, k+1 target character is combined, obtains target license plate number.
Specifically, according to the sequence according to k+1 preset range, k+1 obtained target character is ranked up, mesh is obtained
Mark license plate number.
For example, in a specific embodiment, there is 7 preset ranges, the corresponding sub- word character set of first preset range
Target character be " Shanghai ", the target character of the corresponding sub- word character set of second preset range is " B ", third preset range
The target character of corresponding sub- word character set is " 2 ", and the target character of the corresponding sub- word character set of the 4th preset range is
" 6 ", the target character of the corresponding sub- word character set of the 5th preset range are " A ", the corresponding sub- word word of the 6th preset range
The target character of symbol collection is " 6 ", and the target character of the corresponding sub- word character set of the 7th preset range is " 3 ", is combined
It is " Shanghai B26A63 " to target license plate number.
In the present embodiment, by the way that successively the character in a-c recognition result is put into character set X, and it is pre- according to k+1
If range, the k cut-off of character set X is determined, and then character is carried out to q character in character set X using k cut-off and is cut
Point, k+1 sub- character set are obtained, the most character of frequency of occurrence obtains k+1 as target character in every sub- character set of acquisition
A target character is combined k+1 target character then according to the sequence of k+1 preset range, obtains the target carriage trade mark
Code avoids so that the character in target license plate image can be obtained without being split to target license plate image according to default
Template to target carriage image carry out caused by partial character can not identify that the not high problem of the character accuracy rate identified mentions
The high accuracy rate and stability of Car license recognition.
On the basis of the corresponding embodiment of Fig. 5, below by a specific embodiment come to being mentioned in step S64
And the every sub- character set of acquisition in the most character of frequency of occurrence as target character, obtain the specific of k+1 target character
Implementation method is described in detail.
Referring to Fig. 6, Fig. 6 shows the specific implementation flow of step S64 provided in an embodiment of the present invention, details are as follows:
S641: the character that frequency of occurrence is most in first sub- character set is obtained, as the first character.
Specifically, the number occurred to each character in first sub- character set counts, most by there is number of words
Character as the first character.
For example, in a specific embodiment, first sub- character set be " I ", " L ", " L ", " L ", " L ", " L ", " L ",
" Shanghai ", " Shanghai ", " U ", " Shanghai ", " U ", " Shanghai ", " Shanghai ", " I " }, the character number occurred is counted, statistical result is
Character " L " occurs 6 times, and character " Shanghai " occurs 5 times, and character " U " occurs twice, and character " I " occurs 2 times, and frequency of occurrence is most
Character " L " be used as the first character.
S642: if the first character is non-Chinese character, the Chinese that frequency of occurrence is most in first sub- character set is obtained
Character is as first aim character.
Specifically, in the actual use of license plate, the first aim character of license plate number is Chinese character, but in neural network
Identification in, due to part sliding image in Chinese character it is incomplete, may arrive causes Chinese character to be identified as capitalization English letter, if walk
When first character obtained in rapid S641 is non-Chinese character, then the Chinese that frequency of occurrence is most in first sub- character set is obtained
Character is as first aim character.
By taking the example in step 641 as an example, statistical result is that character " L " occurs 6 times, and character " Shanghai " occurs 5 times, character
" U " occurs twice, and character " I " occurs 2 times, the most character of frequency of occurrence " L " is regard as the first character, by the first character
It is detected, " L " is non-Chinese character, thus, it regard the most Chinese character of the number of appearance " Shanghai " as first aim word
Symbol.
S643: if the first character is Chinese character, using the first character as first aim character.
Specifically, if the first character obtained in step S641 is Chinese character, using first character as first
Target character.
It should be noted that the successive of certainty does not execute sequence by step S642 and step S643, it can be and hold side by side
Capable relationship, herein with no restrictions.
S644: the character that frequency of occurrence is most in j-th of character set is obtained, as j-th of target character, wherein j is greater than
1 and be less than or equal to k+1.
Specifically, for j-th of sub- character set, the character number occurred in the sub- character set is counted, is obtained out
Target character of the most character of occurrence number as the sub- character set.
For example, in a specific embodiment, target character number is 7, i.e. k=6, for the 2nd to the 7th sub- word
Symbol collects, and the most character of frequency of occurrence obtains 6 target characters and be respectively as follows: as target character in every sub- character set of acquisition
" B ", " 2 ", " 3 ", " D ", " 6 " and " 3 ".
In the present embodiment, the most character of frequency of occurrence in first sub- character set is obtained, as the first character, and to the
Whether one character is that Chinese character is judged, if the first character is non-Chinese character, obtains in first sub- character set
The most Chinese character of occurrence number is as first aim character, if the first character is Chinese character, using the first character as
First aim character obtains the character that frequency of occurrence is most in the character set for other characters of non-first character, as
The target character of the character set improves Car license recognition to quick and precisely get each target character of license plate number
Accuracy rate.
It should be understood that the size of the serial number of each step is not meant that the order of the execution order in above-described embodiment, each process
Execution sequence should be determined by its function and internal logic, the implementation process without coping with the embodiment of the present invention constitutes any limit
It is fixed.
In one embodiment, a kind of license plate recognition device is provided, license plate is known in the license plate recognition device and above-described embodiment
Other method corresponds.As shown in fig. 7, the license plate recognition device includes obtaining module 10, preprocessing module 20, sliding block
30, module 40, identification module 50 and segmentation module 60 are terminated.Detailed description are as follows for each functional module:
Module 10 is obtained, for obtaining initial license plate image；
Preprocessing module 20 obtains the target that size is a × b pixel for pre-processing to initial license plate image
License plate image, wherein a, b are positive integer；
Sliding block 30, for using default size for the sliding window of c × b pixel, using 1 pixel as step-length,
Slide is carried out on target license plate image from left to right, is slided every time, the target license plate image within the scope of sliding window is obtained,
As a sliding window image, wherein c < a and c are positive integer；
Module 40 is terminated, after sliding a-c times, slide is terminated, obtains a-c sliding window image；
Identification module 50 identifies for a-c sliding window image to be input in convolutional neural networks, obtains a-
C recognition result, wherein each recognition result includes several characters；
Divide module 60, for all characters in a-c recognition result to be formed character string, according to default partitioning scheme
Character string is split, obtains target character from obtained each substring, and target carriage is determined according to target character
Trade mark code.
Further, preprocessing module 20 includes:
Detection unit 21, for obtaining the coboundary of license plate and license plate in initial license plate image by edge detection algorithm
Lower boundary；
Determination unit 22, for determining the range image of license plate according to the coboundary of license plate and the lower boundary of license plate；
Unit 23 is corrected, for carrying out slant correction to range image using Radon transform, the foundation drawing after being corrected
Picture；
Unit 24 is cut, for being cut centered on the center of gravity of base image to base image, obtaining size is a
The target license plate image of × b pixel.
Further, detection unit 21 includes:
Subelement 211 is denoised, for carrying out noise removal to initial license plate image by Gaussian Blur, obtains denoising license plate
Image；
Computation subunit 212, for calculating the horizontal direction of denoising license plate image using preset gradient operator and hanging down
Histogram to gradient value, obtain Initial Gradient value set；
Subelement 213 is refined, for carrying out edge thinning to Initial Gradient value set by the way of the inhibition of non-maximum value
Processing, obtains the wide gradient edge of pixel；
It filters subelement 214 and obtains gradient side for using the weak marginal point in preset dual threshold filter gradient edge
Strong edge point in edge；
Subelement 215 is delimited, for determining the coboundary of license plate and the lower boundary of license plate according to strong edge point.
Further, segmentation module 60 includes:
Generation unit 61 obtains X={ x for successively the character in a-c recognition result to be put into character set X1,
x2..., xq, wherein q is the sum for the character for including, x in a-c recognition resultiFor i-th of character in character set X, i
∈ [1, q], q are positive integer；
Selection unit 62, for determining the k cut-off of character set X, wherein k < q, and k according to k+1 preset range
For positive integer；
Cutting unit 63 obtains k+1 for using k cut-off to carry out character cutting to q character in character set X
A sub- character set；
Preferred cell 64 obtains k for obtaining frequency of occurrence is most in every sub- character set character as target character
+ 1 target character；
Assembled unit 65 is combined k+1 target character, obtains mesh for the sequence according to k+1 preset range
Mark license plate number.
Further, it is preferable to which unit 64 includes:
Subelement 641 is counted, for obtaining the character that frequency of occurrence is most in first sub- character set, as the first word
Symbol；
First screening subelement 642 obtains in first sub- character set if being non-Chinese character for the first character
The most Chinese character of occurrence number is as first aim character；
Second deletes and selects subelement 643, if being Chinese character for the first character, using the first character as first aim
Character；
Subelement 644 is determined, for obtaining the character that frequency of occurrence is most in j-th of character set, as j-th of target word
Symbol, wherein j is greater than 1 and is less than or equal to k+1.
Specific about license plate recognition device limits the restriction that may refer to above for licence plate recognition method, herein not
It repeats again.Modules in above-mentioned license plate recognition device can be realized fully or partially through software, hardware and combinations thereof.On
Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form
In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 8.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is used to store the initial license plate image of trained convolutional neural networks model and input.The computer is set
Standby network interface is used to communicate with external terminal by network connection.To realize when the computer program is executed by processor
A kind of licence plate recognition method.
In one embodiment, a kind of computer equipment is provided, including memory, processor and storage are on a memory
And the computer program that can be run on a processor, processor realize above-described embodiment Car license recognition side when executing computer program
The step of method, such as step S10 shown in Fig. 2 to step S60.Alternatively, processor realizes above-mentioned reality when executing computer program
Apply the function of each module/unit of license plate recognition device in example, such as module 10 shown in Fig. 7 is to the function of module 60.To avoid
It repeats, which is not described herein again.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Licence plate recognition method in above method embodiment is realized when machine program is executed by processor, alternatively, the computer program is processed
The function of each module/unit in license plate recognition device in above-mentioned apparatus embodiment is realized when device executes.To avoid repeating, here not
It repeats again.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
It is apparent to those skilled in the art that for convenience of description and succinctly, only with above-mentioned each function
Can unit, module division progress for example, in practical application, can according to need and by above-mentioned function distribution by different
Functional unit, module are completed, i.e., the internal structure of described device is divided into different functional unit or module, more than completing
The all or part of function of description.
Embodiment described above is merely illustrative of the technical solution of the present invention, rather than its limitations；Although referring to aforementioned reality
Applying example, invention is explained in detail, those skilled in the art should understand that: it still can be to aforementioned each
Technical solution documented by embodiment is modified or equivalent replacement of some of the technical features；And these are modified
Or replacement, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution should all
It is included within protection scope of the present invention.
Claims (10)
1. a kind of licence plate recognition method, which is characterized in that the licence plate recognition method includes:
Obtain initial license plate image；
The initial license plate image is pre-processed, obtains the target license plate image that size is a × b pixel, wherein a, b
For positive integer；
Use default size for the sliding window of c × b pixel, using 1 pixel as step-length, on the target license plate image from
Left-to-right carries out slide, slides every time, a target license plate image within the scope of the sliding window is obtained, as one
Sliding window image, wherein c < a and c are positive integer；
After sliding a-c times, slide is terminated, a-c sliding window image is obtained；
The a-c sliding window images are input in convolutional neural networks and are identified, a-c recognition result is obtained,
In, each recognition result includes several characters；
All characters in a-c recognition results are formed into character string, according to default partitioning scheme to the character string into
Row segmentation, obtains target character, and determine target license plate number according to the target character from obtained each substring.
2. licence plate recognition method as described in claim 1, which is characterized in that described to be located in advance to the initial license plate image
Reason, obtaining the target license plate image that size is a × b pixel includes:
By edge detection algorithm, the coboundary of license plate and the lower boundary of license plate in the initial license plate image are obtained；
According to the lower boundary of the coboundary of the license plate and the license plate, the range image of license plate is determined；
Slant correction is carried out to the range image using Radon transform, the base image after being corrected；
Centered on the center of gravity of the base image, the base image is cut, obtaining size is a × b pixel
The target license plate image.
3. licence plate recognition method as claimed in claim 2, which is characterized in that it is described by edge detection algorithm, described in acquisition
The coboundary of license plate and the lower boundary of license plate include: in initial license plate image
Noise removal is carried out to the initial license plate image by Gaussian Blur, obtains denoising license plate image；
The gradient value horizontally and vertically that the denoising license plate image is calculated using preset gradient operator, is obtained
Initial Gradient value set；
Edge thinning processing is carried out to the Initial Gradient value set by the way of the inhibition of non-maximum value, it is wide to obtain a pixel
Gradient edge；
The weak marginal point in the gradient edge is filtered using preset dual threshold, obtains the strong edge in the gradient edge
Point；
According to the strong edge point, the coboundary of the license plate and the lower boundary of the license plate are determined.
4. licence plate recognition method as described in any one of claims 1 to 3, which is characterized in that described to tie the a-c identifications
All characters in fruit form character string, and obtain target character from the character string according to predetermined manner, according to the mesh
Marking-up, which accords with, determines that target license plate number includes:
Successively all characters in the a-c recognition results are put into character set X, obtain X={ x1, x2..., xq,
In, q is the sum for the character for including, x in a-c recognition resultiFor i-th of character in character set X, i ∈ [1, q], q is positive
Integer；
According to k+1 preset range, the k cut-off of the character set X is determined, wherein k < q, and k is positive integer；
Character cutting is carried out to q character in the character set X using the k cut-offs, obtains k+1 sub- character set；
Frequency of occurrence is most in each sub- character set character is obtained as target character, obtains the k+1 target words
Symbol；
According to the sequence of the k+1 preset ranges, the k+1 target characters are combined, the target license plate is obtained
Number.
5. licence plate recognition method as claimed in claim 4, which is characterized in that described obtain occurs in each sub- character set
The most character of number includes: as target character
The character that frequency of occurrence is most in first sub- character set is obtained, as the first character；
If first character is non-Chinese character, the middle text that frequency of occurrence is most in first sub- character set is obtained
Symbol is used as first target character；
If first character is Chinese character, using first character as first target character；
The character that frequency of occurrence is most in j-th of character set is obtained, as j-th of target character, wherein j is greater than 1
And it is less than or equal to k+1.
6. a kind of license plate recognition device, which is characterized in that the license plate recognition device includes:
Module is obtained, for obtaining initial license plate image；
Preprocessing module obtains the target carriage that size is a × b pixel for pre-processing to the initial license plate image
Board image, wherein a, b are positive integer；
Sliding block, for using default size for the sliding window of c × b pixel, using 1 pixel as step-length, in the mesh
Slide is carried out from left to right on mark license plate image, is slided every time, is obtained the target license plate figure within the scope of the sliding window
Picture, as a sliding window image, wherein c < a and c are positive integer；
Module is terminated, after sliding a-c times, slide is terminated, obtains a-c sliding window image；
Identification module identifies for the a-c sliding window images to be input in convolutional neural networks, obtains a-c
A recognition result, wherein each recognition result includes several characters；
Divide module, for the character in the a-c recognition results to be formed character string, according to default partitioning scheme to described
Character string is split, and target character is obtained from obtained each substring, and determine target according to the target character
License plate number.
7. license plate recognition device as claimed in claim 6, which is characterized in that the preprocessing module includes:
Detection unit, for by edge detection algorithm, obtaining the coboundary of license plate and license plate in the initial license plate image
Lower boundary；
Determination unit, for determining the range image of license plate according to the coboundary of the license plate and the lower boundary of the license plate；
Unit is corrected, for carrying out slant correction to the range image using Radon transform, the base image after being corrected；
Unit is cut, for being cut centered on the center of gravity of the base image to the base image, obtaining size is
The target license plate image of a × b pixel.
8. license plate recognition device as claimed in claim 6, which is characterized in that the segmentation module includes:
Generation unit obtains X={ x for successively the character in the a-c recognition results to be put into character set X1,
x2..., xq, wherein q is the sum for the character for including, x in a-c recognition resultiFor i-th of character in character set X, i
∈ [1, q], q are positive integer；
Selection unit, for determining the k cut-off of the character set X according to k+1 preset range, wherein k < q, and k is
Positive integer；
Cutting unit obtains k+ for using the k cut-offs to carry out character cutting to q character in the character set X
1 sub- character set；
Preferred cell obtains k+1 for obtaining frequency of occurrence is most in each sub- character set character as target character
A target character；
Assembled unit is combined the k+1 target characters, obtains for the sequence according to the k+1 preset ranges
The target license plate number.
9. a kind of computer equipment, including memory, processor and storage are in the memory and can be in the processor
The computer program of upper operation, which is characterized in that the processor realized when executing the computer program as claim 1 to
The step of any one of 5 licence plate recognition method.
10. a kind of computer readable storage medium, the computer-readable recording medium storage has computer program, and feature exists
In the step of realization licence plate recognition method as described in any one of claim 1 to 5 when the computer program is executed by processor
Suddenly.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811113599.4A CN109492642B (en) | 2018-09-25 | 2018-09-25 | License plate recognition method, license plate recognition device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811113599.4A CN109492642B (en) | 2018-09-25 | 2018-09-25 | License plate recognition method, license plate recognition device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109492642A true CN109492642A (en) | 2019-03-19 |
| CN109492642B CN109492642B (en) | 2023-11-24 |
Family
ID=65689336
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811113599.4A Active CN109492642B (en) | 2018-09-25 | 2018-09-25 | License plate recognition method, license plate recognition device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109492642B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232381A (en) * | 2019-06-19 | 2019-09-13 | 梧州学院 | License Plate Segmentation method, apparatus, computer equipment and computer readable storage medium |
| CN110264489A (en) * | 2019-06-24 | 2019-09-20 | 北京奇艺世纪科技有限公司 | A kind of image boundary detection method, device and terminal |
| CN111144390A (en) * | 2019-12-20 | 2020-05-12 | 武汉瑞纳捷电子技术有限公司 | Missing character searching method and system in license plate recognition and electronic equipment |
| CN111428708A (en) * | 2020-02-26 | 2020-07-17 | 深圳壹账通智能科技有限公司 | License plate checking method and device, computer equipment and storage medium |
| CN111681229A (en) * | 2020-06-10 | 2020-09-18 | 创新奇智(上海)科技有限公司 | Deep learning model training method, wearable clothes flaw identification method and wearable clothes flaw identification device |
| CN112434700A (en) * | 2020-11-25 | 2021-03-02 | 创新奇智(上海)科技有限公司 | License plate recognition method, device, equipment and storage medium |
| CN112686246A (en) * | 2019-10-18 | 2021-04-20 | 深圳市优必选科技股份有限公司 | License plate character segmentation method and device, storage medium and terminal equipment |
| CN112906643A (en) * | 2021-03-22 | 2021-06-04 | 高新兴科技集团股份有限公司 | License plate number identification method and device |
| CN113688927A (en) * | 2021-08-31 | 2021-11-23 | 中国平安人寿保险股份有限公司 | Picture sample generation method and device, computer equipment and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5542006A (en) * | 1994-06-21 | 1996-07-30 | Eastman Kodak Company | Neural network based character position detector for use in optical character recognition |
| CN1801181A (en) * | 2006-01-06 | 2006-07-12 | 华南理工大学 | Robot capable of automatically recognizing face and vehicle license plate |
| US8761498B1 (en) * | 2012-01-26 | 2014-06-24 | Google Inc. | Face and license plate detection in street level images with 3-D road width features estimated from laser data |
| CN104715252A (en) * | 2015-03-12 | 2015-06-17 | 电子科技大学 | License plate character segmentation method with combination of dynamic template and pixel points |
| CN105373794A (en) * | 2015-12-14 | 2016-03-02 | 河北工业大学 | Vehicle license plate recognition method |
| CN107016366A (en) * | 2017-03-29 | 2017-08-04 | 浙江师范大学 | A kind of guideboard detection method based on Adaptive windowing mouthful and convolutional neural networks |
| CN108416348A (en) * | 2018-01-29 | 2018-08-17 | 重庆邮电大学 | Plate location recognition method based on support vector machines and convolutional neural networks |
-
2018
- 2018-09-25 CN CN201811113599.4A patent/CN109492642B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5542006A (en) * | 1994-06-21 | 1996-07-30 | Eastman Kodak Company | Neural network based character position detector for use in optical character recognition |
| CN1801181A (en) * | 2006-01-06 | 2006-07-12 | 华南理工大学 | Robot capable of automatically recognizing face and vehicle license plate |
| US8761498B1 (en) * | 2012-01-26 | 2014-06-24 | Google Inc. | Face and license plate detection in street level images with 3-D road width features estimated from laser data |
| CN104715252A (en) * | 2015-03-12 | 2015-06-17 | 电子科技大学 | License plate character segmentation method with combination of dynamic template and pixel points |
| CN105373794A (en) * | 2015-12-14 | 2016-03-02 | 河北工业大学 | Vehicle license plate recognition method |
| CN107016366A (en) * | 2017-03-29 | 2017-08-04 | 浙江师范大学 | A kind of guideboard detection method based on Adaptive windowing mouthful and convolutional neural networks |
| CN108416348A (en) * | 2018-01-29 | 2018-08-17 | 重庆邮电大学 | Plate location recognition method based on support vector machines and convolutional neural networks |
Non-Patent Citations (1)
| Title |
|---|
| 邱刚 等: "基于边缘特征和神经网络的汽车牌照定位算法", 微机发展, vol. 15, no. 04, pages 30 - 32 * |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232381A (en) * | 2019-06-19 | 2019-09-13 | 梧州学院 | License Plate Segmentation method, apparatus, computer equipment and computer readable storage medium |
| CN110264489A (en) * | 2019-06-24 | 2019-09-20 | 北京奇艺世纪科技有限公司 | A kind of image boundary detection method, device and terminal |
| CN110264489B (en) * | 2019-06-24 | 2022-07-05 | 北京奇艺世纪科技有限公司 | Image boundary detection method and device and terminal |
| CN112686246A (en) * | 2019-10-18 | 2021-04-20 | 深圳市优必选科技股份有限公司 | License plate character segmentation method and device, storage medium and terminal equipment |
| CN112686246B (en) * | 2019-10-18 | 2024-01-02 | 深圳市优必选科技股份有限公司 | License plate character segmentation method and device, storage medium and terminal equipment |
| CN111144390A (en) * | 2019-12-20 | 2020-05-12 | 武汉瑞纳捷电子技术有限公司 | Missing character searching method and system in license plate recognition and electronic equipment |
| CN111144390B (en) * | 2019-12-20 | 2023-10-27 | 武汉瑞纳捷半导体有限公司 | Missing character searching method and system in license plate recognition and electronic equipment |
| CN111428708A (en) * | 2020-02-26 | 2020-07-17 | 深圳壹账通智能科技有限公司 | License plate checking method and device, computer equipment and storage medium |
| CN111681229A (en) * | 2020-06-10 | 2020-09-18 | 创新奇智(上海)科技有限公司 | Deep learning model training method, wearable clothes flaw identification method and wearable clothes flaw identification device |
| CN112434700A (en) * | 2020-11-25 | 2021-03-02 | 创新奇智(上海)科技有限公司 | License plate recognition method, device, equipment and storage medium |
| CN112906643A (en) * | 2021-03-22 | 2021-06-04 | 高新兴科技集团股份有限公司 | License plate number identification method and device |
| CN113688927A (en) * | 2021-08-31 | 2021-11-23 | 中国平安人寿保险股份有限公司 | Picture sample generation method and device, computer equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109492642B (en) | 2023-11-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109492642A (en) | Licence plate recognition method, device, computer equipment and storage medium | |
| CN110008809B (en) | Method and device for acquiring form data and server | |
| CN109472262A (en) | Licence plate recognition method, device, computer equipment and storage medium | |
| CN107945185B (en) | Image segmentation method and system based on wide residual pyramid pooling network | |
| CN109447117A (en) | The double-deck licence plate recognition method, device, computer equipment and storage medium | |
| CN110163842B (en) | Building crack detection method and device, computer equipment and storage medium | |
| CN103077388B (en) | Fast text towards portable computing device sweeps the method for reading | |
| CN104794685B (en) | A kind of method and device for realizing image denoising | |
| CN109670500A (en) | A kind of character area acquisition methods, device, storage medium and terminal device | |
| CN109635656A (en) | Vehicle attribute recognition methods, device, equipment and medium neural network based | |
| US11790499B2 (en) | Certificate image extraction method and terminal device | |
| CN111680690B (en) | Character recognition method and device | |
| CN109543548A (en) | A kind of face identification method, device and storage medium | |
| CN108280829A (en) | Welding seam method, computer installation and computer readable storage medium | |
| WO2019056346A1 (en) | Method and device for correcting tilted text image using expansion method | |
| CN111951154B (en) | Picture generation method and device containing background and medium | |
| CN111353491B (en) | Text direction determining method, device, equipment and storage medium | |
| EP2613294A1 (en) | System and method for synthesizing portrait sketch from photo | |
| CN110942071A (en) | License plate recognition method based on license plate classification and LSTM | |
| CN109858409A (en) | Manual figure conversion method, device, equipment and medium | |
| CN111159150A (en) | Data expansion method and device | |
| CN105825228A (en) | Image identification method and apparatus | |
| Swaileh et al. | Multi-script iterative steerable directional filtering for handwritten text line extraction | |
| CN103455816B (en) | Stroke width extraction method and device and character recognition method and system | |
| CN113850238A (en) | Document detection method and device, electronic equipment and storage medium |
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 |