CN106203418A - A kind of method and device of car plate detection - Google Patents
A kind of method and device of car plate detection Download PDFInfo
- Publication number
- CN106203418A CN106203418A CN201610556058.3A CN201610556058A CN106203418A CN 106203418 A CN106203418 A CN 106203418A CN 201610556058 A CN201610556058 A CN 201610556058A CN 106203418 A CN106203418 A CN 106203418A
- Authority
- CN
- China
- Prior art keywords
- car plate
- character
- information
- plate
- zone position
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06V10/225—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 a marking or identifier characterising the area
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
-
- 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
- G06V30/146—Aligning or centring of the image pick-up or image-field
- G06V30/1475—Inclination or skew detection or correction of characters or of image to be recognised
- G06V30/1478—Inclination or skew detection or correction of characters or of image to be recognised of characters or characters lines
-
- 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
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
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)
- Character Input (AREA)
Abstract
The method that the invention discloses car plate detection, including: obtain the image information that image capture device gathers;Utilize character machining device that image information is detected, obtain the zone position information of each character in image information and the character kind of information of correspondence;Zone position information according to each character and the character kind of information of correspondence and car plate rule are fitted obtaining matching car plate to the car plate in image information, and matching car plate is carried out License Plate obtain License Plate result;According to License Plate result, car plate is carried out angle correct;Car plate after correction is carried out character recognition, obtains license board information;The method uses first to correct and identifies afterwards, reduces angle and tilts the impact on recognizer, increases the discrimination of Car license recognition;And take full advantage of car plate spatial positional information and character kind of information, strengthen the algorithm recall rate to car plate, it is to avoid flase drop;The invention also discloses the device of car plate detection, there is the effect above.
Description
Technical field
The present invention relates to image identification technical field, the method and device detected particularly to a kind of car plate.
Background technology
Modern society, vehicle market recoverable amount is the hugest, and also can continue to increase along with China's expanding economy.With
And come is also to represent, to vehicle, information that traffic is relevant, the situation quickly increased.Intelligent transportation system is as current science and technology
The highest achievement, it presents powerful advantage at the management aspect of information of vehicles.And car plate automatization identifies and hands over as intelligence
Most important link in way system, has directly influenced the use of intelligent transportation system.
Currently, license plate recognition technology utilization rate based on video is more and more higher, but for vehicle during road-surface concrete and shooting
The situation utilization rate that the visual angle of machine there are differences is the highest, because this mode is compared with tradition bayonet socket scene monitoring, and vehicle mesh
Rotating savings has the inclination of various angle.This makes tradition detector model based on horizontal vertical texture information cannot adapt to multiple regarding
The vehicle detection at angle, especially the detection identification to high inclination-angle target, its recall rate is the lowest.Therefore, how to adapt to multiple visual angle
The detection of vehicle license plate, be those skilled in the art's technical issues that need to address.
Summary of the invention
It is an object of the invention to provide the method and device of a kind of car plate detection, utilization first corrects and identifies afterwards, reduces angle
Degree tilts the impact on recognizer, increases the discrimination of Car license recognition;Make full use of car plate spatial positional information and character
Kind of information, strengthens the algorithm recall rate to car plate, it is to avoid flase drop.
For solving above-mentioned technical problem, the present invention provides a kind of method that car plate detects, including:
Obtain the image information that image capture device gathers;
Utilize character machining device that described image information is detected, obtain the position, region of each character in described image information
Confidence breath and the character kind of information of correspondence;
Zone position information according to each described character and the character kind of information of correspondence and car plate rule are to described image
Car plate in information is fitted obtaining matching car plate, and described matching car plate carries out License Plate obtains License Plate knot
Really;
According to described License Plate result, described car plate is carried out angle correct;
Described car plate after correction is carried out character recognition, obtains described license board information.
Wherein, regular to described according to the zone position information of each described character and the character kind of information of correspondence and car plate
Car plate in image information is fitted obtaining matching car plate, including:
Zone position information according to each described character and the character kind of information of correspondence, determine chinese character and correspondence
The zone position information of described chinese character;
Whether alphabetic character is there is in the range of the zone position information preset distance of chinese character described in judging distance;
If existing, then according to the zone position information of described chinese character and the zone position information of the alphabetic character of correspondence
Determine predetermined car plate direction;
Described predetermined car plate direction determines the zone position of the alphabetic character of remaining five alphabetic character and correspondence
Breath, obtains meeting the matching car plate of seven characters of car plate rule.
Wherein, described matching car plate is carried out License Plate and obtains License Plate result, including:
According to the zone position information of each character in described matching car plate, ask for the boundary rectangle of described matching car plate;
Described boundary rectangle is enlarged according to predetermined ratio, obtains License Plate result.
Wherein, according to described License Plate result, described car plate is carried out angle correct, including:
Determine the central point of the zone position information of each character in described matching car plate, fit to according to each described central point
First straight line, and the level inclination of car plate is determined according to described first straight line and horizontal angle;
Calculate the vertical dip angle of each character in described matching car plate, and will be according to the calculated meansigma methods of each vertical dip angle
Vertical dip angle as car plate;
According to described level inclination and described vertical dip angle, utilize affine transformation that described car plate is carried out angle correct.
Wherein, the training method of described character machining device, including:
Obtain the character data of various different angles as nominal data;
Obtain multiple character image datas of various different visual angles as training sample;
According to described nominal data, utilize CNN convolutional neural networks model that described training sample is trained, obtain word
Symbol detector.
The present invention also provides for the device of a kind of car plate detection, including:
Acquisition module, for obtaining the image information that image capture device gathers;
Character machining module, is used for utilizing character machining device to detect described image information, obtains described image letter
The zone position information of each character and the character kind of information of correspondence in breath;
Car plate matching locating module, for the zone position information according to each described character and the character kind of information of correspondence
With car plate rule, the car plate in described image information is fitted obtaining matching car plate, and described matching car plate is carried out car plate
Location obtains License Plate result;
VLP correction module, for carrying out angle correct according to described License Plate result to described car plate;
Recognition of License Plate Characters module, for the described car plate after correction is carried out character recognition, obtains described license board information.
Wherein, described car plate matching locating module includes car plate fitting unit, and wherein, described car plate fitting unit includes:
Chinese character subelement, for the zone position information according to each described character and the character kind of information of correspondence,
Determine the zone position information of the described chinese character of chinese character and correspondence;
First alphabetic character judgment sub-unit, for the zone position information preset distance of chinese character described in judging distance
In the range of whether there is alphabetic character;
Car plate direction determines subelement, if for existing, then according to zone position information and the correspondence of described chinese character
The zone position information of alphabetic character determine predetermined car plate direction;
Car plate matching subelement, for determining remaining five alphabetic characters and word of correspondence on described predetermined car plate direction
The zone position information of alphabetic character, obtains meeting the matching car plate of seven characters of car plate rule.
Wherein, described car plate matching locating module includes License Plate unit, and wherein, described License Plate unit includes:
Boundary rectangle obtains subelement, for according to the zone position information of each character in described matching car plate, asks for institute
State the boundary rectangle of matching car plate;
License Plate subelement, for being enlarged according to predetermined ratio by described boundary rectangle, obtains License Plate knot
Really.
Wherein, described VLP correction module, including:
Level inclination unit, for determining the central point of the zone position information of each character in described matching car plate, according to
Each described central point fits to the first straight line, and determines that the level of car plate is inclined according to described first straight line and horizontal angle
Angle;
Vertical dip angle unit, for calculating the vertical dip angle of each character in described matching car plate, and vertically will incline according to each
The calculated meansigma methods in angle is as the vertical dip angle of car plate;
VLP correction unit, for according to described level inclination and described vertical dip angle, utilizes affine transformation to described car
Board carries out angle correct.
Wherein, also include: training module, for described character machining device is trained;Specifically include:
Nominal data unit, for obtaining the character data of various different angles as nominal data;
Training sample unit, for obtaining multiple character image datas of various different visual angles as training sample;
Training unit, for according to described nominal data, utilizes CNN convolutional neural networks model to enter described training sample
Row training, obtains character machining device.
The method of car plate provided by the present invention detection, including: obtain the image information that image capture device gathers;Utilize
Image information is detected by character machining device, obtains the zone position information of each character in image information and the character kind of correspondence
Category information;Zone position information according to each character and the character kind of information of correspondence and car plate rule are to the car in image information
Board is fitted obtaining matching car plate, and matching car plate is carried out License Plate obtains License Plate result;According to License Plate
Result carries out angle correct to car plate;Car plate after correction is carried out character recognition, obtains license board information;
The method uses first to correct and identifies afterwards, reduces angle and tilts the impact on recognizer, increases Car license recognition
Discrimination;And take full advantage of car plate spatial positional information and character kind of information carries out car plate matching and License Plate, increase
The strong algorithms recall rate to car plate, it is to avoid flase drop;Adapt to the detection of the vehicle license plate to multiple visual angle;Solve to examine on road
When surveying license board information, because the restriction of camera position, there is the biggest angle of inclination in the car plate of collection, causes detection vehicle car
The problem that board information is failed;The invention also discloses the device of car plate detection, there is the effect above, do not repeat them here.
Accompanying drawing explanation
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
In having technology to describe, the required accompanying drawing used is briefly described, it should be apparent that, the accompanying drawing in describing below is only this
Inventive embodiment, for those of ordinary skill in the art, on the premise of not paying creative work, it is also possible to according to
The accompanying drawing provided obtains other accompanying drawing.
The flow chart of the method for the car plate detection that Fig. 1 is provided by the embodiment of the present invention;
The schematic diagram of the vertical dip angle of the car plate that Fig. 2 is provided by the embodiment of the present invention;
The flow process signal of the method for a kind of concrete oblique viewing angle car plate detection that Fig. 3 is provided by the embodiment of the present invention
Figure;
The structured flowchart of the device of the car plate detection that Fig. 4 is provided by the embodiment of the present invention.
Detailed description of the invention
The core of the present invention is to provide the method and device of a kind of car plate detection, and utilization first corrects and identifies afterwards, reduces angle
Degree tilts the impact on recognizer, increases the discrimination of Car license recognition;Make full use of car plate spatial positional information and character
Kind of information, strengthens the algorithm recall rate to car plate, it is to avoid flase drop.
For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely described, it is clear that described embodiment is
The a part of embodiment of the present invention rather than whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under not making creative work premise, broadly falls into the scope of protection of the invention.
Refer to the flow chart of the method for the car plate detection that Fig. 1, Fig. 1 are provided by the embodiment of the present invention;The method is permissible
Including:
The image information that S100, acquisition image capture device gather;
Wherein, the method for this car plate detection may be used for the process of the vehicle image captured, it is also possible to for regarding of gathering
Frequently the process of two field picture;The most here image information can be video image information can also be picture image information.Herein
The concrete form of image information and image capture device is not defined.
S110, utilize character machining device that described image information is detected, obtain each character in described image information
The character kind of information of zone position information and correspondence;
Wherein, this step, primarily to detect the character in image information, obtains each in image information
The zone position information of character and the character kind of information of correspondence;The such as characters on license plate of China can be divided into three kinds, Chinese Character
Symbol, alphabetic character and numerical character.All characters in image can be detected during whole image information detects,
Detection can obtain the zone position information of character and the kind of information of character.
Herein for improving further to various characters especially recall rate to wide-angle character, it is possible to use each class
Character machining device is trained by the training sample data of the magnanimity of oblique viewing angle.But the not tool to training character machining device
Body Model is defined;Preferably, the training method of character machining device, including:
Obtain the character data of various different angles as nominal data;
Obtain multiple character image datas of various different visual angles as training sample;
According to described nominal data, utilize CNN convolutional neural networks model that described training sample is trained, obtain word
Symbol detector.
Wherein, character machining device learns CNN convolutional neural networks model construction according to the degree of depth, can utilize each during training
Plant the mass data of visual angle kinds of characters as training sample.Because each class oblique viewing angle has the number of training of magnanimity
According to, the nominal data of utilization has the character data of various different angles, so detection model i.e. tilts word to wide-angle character
Symbol has good recall rate.
When utilizing character machining device that image information is detected, if image failing find character information, then under carrying out
One frame analysis, until spreading out of number and the correspondence position information of character after finding character information.
S120, according to the zone position information of each described character and the character kind of information of correspondence and car plate rule to described
Car plate in image information is fitted obtaining matching car plate, and described matching car plate is carried out License Plate obtains License Plate
Result;
Wherein, here according to the zone position information of each character and the character kind of information of correspondence and car plate rule to image
When car plate in information is fitted and positions, to be fitted with specific reference to the car plate rule that every country is concrete;Here
The character quantity of car plate of rule various countries to be comprised, character kind and the queueing discipline of character.According to such arrangement rule
Then carry out the character location information obtained and kind of information selecting until being determined for compliance with a string character letter of all car plates rule
A string character information that breath will determine is as the license board information of matching, it is possible to utilize the position of a string character information determined
Information carries out License Plate.
Concrete process may is that
From the zone position information of each described character and the character kind of information of correspondence, first is determined according to car plate rule
Spcial character;
The second spcial character is determined in the preset range of the first spcial character described in distance determined by;
In the preset range of the front character once determined, character late is determined successively, until obtaining according to car plate rule
Meet the matching car plate of car plate rule;
Or owing to characters on license plate carries out horizontally arranged according to a direction, accordingly it is also possible to special according to first
Character and the second spcial character determine predetermined car plate direction, carry out the position kind of other characters in the range of the direction determines
Determine, thus obtain matching car plate.
Such as comprise 7 characters at China Coach car plate, and first character is the Chinese character representing province, second character
For alphabetic character, remaining five is alphabetic character or numerical character;It is preferred, therefore, that according to the position, region of each described character
Car plate in described image information is fitted obtaining matching car by confidence breath and the character kind of information of correspondence and car plate rule
Board, including:
Zone position information according to each described character and the character kind of information of correspondence, determine chinese character and correspondence
The zone position information of described chinese character;
Whether alphabetic character is there is in the range of the zone position information preset distance of chinese character described in judging distance;
If existing, then according to the zone position information of described chinese character and the zone position information of the alphabetic character of correspondence
Determine predetermined car plate direction;
Described predetermined car plate direction determines the zone position of the alphabetic character of remaining five alphabetic character and correspondence
Breath, obtains meeting the matching car plate of seven characters of car plate rule.
Wherein, the occurrence of preset distance according to distance between characters on license plate in practical situation and can incline various
In the case of Xie, the actual value of this distance is determined.Here on predetermined car plate direction, determine remaining five alphabetic characters and correspondence
The zone position information of alphabetic character time can be on the line direction from chinese character to alphabetic character.
Wherein, when chinese character more than one, can proceed as described above successively, it is also possible to carry out above-mentioned mistake side by side
Journey.
Detailed process may is that
So when car plate matching, the first step is location chinese character, searches out Chinese Character in the most all characters
Symbol, if can not find, directly carrying out next frame analysis, if finding chinese character, at search word alphabetic character about, determining
May determine that the general orientation of car plate after the two character, look for remaining five characters in corresponding orientation.To each
Chinese character performs above operation, repeats this process n times, completes all car plate matchings.So complete seven characters on license plate
Matching and the location of each character.
I.e. from the beginning of a Chinese character position, finding a neighbouring English alphabet about.According to Chinese character and letter
Seat relatively speculates the position of character late, and searches in character machining result.Successively according to it has been determined that character pre-
Survey the position of character late.If 5 letter or numbers can be found continuously, the most successfully simulate a car plate.
Next need that described matching car plate is carried out License Plate and obtain License Plate result, every according to matching car plate
The zone position information of individual character may determine that the zone position information of the position shared by this car plate, in order to ensure that car plate position is believed
Cease completely can also carry out a certain degree of expansion as License Plate result using the zone position information of this shared position.
Preferably, described matching car plate is carried out License Plate and obtains License Plate result, including:
According to the zone position information of each character in described matching car plate, ask for the boundary rectangle of described matching car plate;
Described boundary rectangle is enlarged according to predetermined ratio, obtains License Plate result.
Wherein, after obtaining meeting the character that car plate is regular, ask for according to the zone position information of character each in matching car plate
Boundary rectangle also extends in proportion to surrounding, obtains License Plate result.
Here predetermined ratio is actually needed according to user and is determined, and does not the most limit concrete proportional numerical value
Fixed.
S130, according to described License Plate result, described car plate is carried out angle correct;
Wherein, the matching car plate obtained through said process is the car plate with certain inclination angle, in order to improve dump car
The recall rate of board and detection effect, to carry out angle correct here, form the matching car plate of a level matching car plate.I.e. root
Car plate and horizontal line and/or the inclination angle of vertical curve is may determine that, then according to these inclination angles to car plate angle according to License Plate result
Degree is corrected.Preferably, according to described License Plate result, described car plate is carried out angle correct, including:
Determine the central point of the zone position information of each character in described matching car plate, fit to according to each described central point
First straight line, and the level inclination of car plate is determined according to described first straight line and horizontal angle;
Calculate the vertical dip angle of each character in described matching car plate, and will be according to the calculated meansigma methods of each vertical dip angle
Vertical dip angle as car plate;
According to described level inclination and described vertical dip angle, utilize affine transformation that described car plate is carried out angle correct.
Wherein, the level inclination of car plate can simulate straight line by the central point of characters on license plate and obtain, car plate
Vertical dip angle by calculate single characters on license plate vertical dip angle, 7 angles can be obtained, then these angles are averaging
Value can be obtained by the vertical dip angle of whole car plate, as shown in Figure 2;Affine change can be utilized according to level angle and vertical angle
Change and car plate is corrected.
S140, to correction after described car plate carry out character recognition, obtain described license board information.
Wherein, characters on license plate can be identified by character recognition here according to OCR model, obtains license plate recognition result
I.e. license board information.
Turn next to Fig. 3, it is provided that a kind of method that concrete oblique viewing angle Car license recognition judges;First have to character is supervised
Survey device to initialize, after having initialized, the frame of video obtained is carried out character machining and classification, obtain according to monitoring result
Chinese character number and positional information, obtain each chinese character of N;Now cross and do not get chinese character next frame to be carried out
The process of video frame images;Obtain a chinese character position, to search word alphabetic character around this chinese character, by this Chinese Character
Symbol and alphabetic character determine car plate direction in space, search for remaining five car plate according to car plate spatial information on the basis of result above
Character, if not finding remaining five character, does not fits to car plate, and utilize second repeat to n-th chinese character position on
Process of stating is until to remaining five characters on license plate;If finding remaining five characters on license plate, then navigate in license plate area and car plate
Each character zone, and according to the license plate sloped angle of character zone position matching, carry out affine transformation, and to the image after conversion
Carry out Recognition of License Plate Characters, export recognition result, complete the detection of oblique viewing angle car plate.
Based on technique scheme, the method for the car plate detection that the embodiment of the present invention provides, after the method uses first correction
Identify, reduce angle and tilt the impact on recognizer, increase the discrimination of Car license recognition;And it is empty to take full advantage of car plate
Between positional information and character kind of information carry out car plate matching and License Plate, strengthen the algorithm recall rate to car plate, it is to avoid
Flase drop;Adapt to the detection of the vehicle license plate to multiple visual angle;When solving to detect license board information on road, because camera position
Restriction, there is the biggest angle of inclination in the car plate of collection, causes the problem that detection vehicle license plate information is failed.
There is the biggest angle of inclination for car plate, cause the problem identifying that license board information is failed, (such as road boundary detection car plate
Time) the method has the advantage that the first employing video analysis mode realizes high inclination-angle Car license recognition, has saved hardware costs,
Reduce labor cost.Second employing is first corrected and is identified afterwards, reduces angle and tilts the impact on recognizer, increases car
The discrimination of board identification.3rd takes full advantage of car plate spatial positional information and character kind of information, strengthens algorithm to car plate
Recall rate, it is to avoid flase drop.4th character machining utilizes the degree of depth to learn CNN convolutional neural networks model, utilizes in the training stage
Kinds of characters data under various visual angles, increase the algorithm recall rate to big tilted character.
The method embodiments providing car plate detection, utilization first corrects and identifies afterwards, reduces angle and tilts knowledge
The impact of other algorithm, increases the discrimination of Car license recognition.
The device of the car plate detection provided the embodiment of the present invention below is introduced, the dress of car plate described below detection
Putting the method detected with above-described car plate can be mutually to should refer to.
Refer to the structured flowchart of the device of the car plate detection that Fig. 4, Fig. 4 are provided by the embodiment of the present invention.This device can
To include:
Acquisition module 100, for obtaining the image information that image capture device gathers;
Character machining module 200, is used for utilizing character machining device to detect described image information, obtains described image
The zone position information of each character and the character kind of information of correspondence in information;
Car plate matching locating module 300, for the zone position information according to each described character and the character kind of correspondence
Information and car plate rule are fitted obtaining matching car plate to the car plate in described image information, and carry out described matching car plate
License Plate obtains License Plate result;
VLP correction module 400, for carrying out angle correct according to described License Plate result to described car plate;
Recognition of License Plate Characters module 500, for the described car plate after correction is carried out character recognition, obtains described car plate letter
Breath.
Optionally, described car plate matching locating module 300 includes car plate fitting unit, wherein, described car plate fitting unit
Including:
Chinese character subelement, for the zone position information according to each described character and the character kind of information of correspondence,
Determine the zone position information of the described chinese character of chinese character and correspondence;
First alphabetic character judgment sub-unit, for the zone position information preset distance of chinese character described in judging distance
In the range of whether there is alphabetic character;
Car plate direction determines subelement, if for existing, then according to zone position information and the correspondence of described chinese character
The zone position information of alphabetic character determine predetermined car plate direction;
Car plate matching subelement, for determining remaining five alphabetic characters and word of correspondence on described predetermined car plate direction
The zone position information of alphabetic character, obtains meeting the matching car plate of seven characters of car plate rule.
Optionally, described car plate matching locating module 300 includes License Plate unit, wherein, described License Plate unit
Including:
Boundary rectangle obtains subelement, for according to the zone position information of each character in described matching car plate, asks for institute
State the boundary rectangle of matching car plate;
License Plate subelement, for being enlarged according to predetermined ratio by described boundary rectangle, obtains License Plate knot
Really.
Optionally, described VLP correction module 400, including:
Level inclination unit, for determining the central point of the zone position information of each character in described matching car plate, according to
Each described central point fits to the first straight line, and determines that the level of car plate is inclined according to described first straight line and horizontal angle
Angle;
Vertical dip angle unit, for calculating the vertical dip angle of each character in described matching car plate, and vertically will incline according to each
The calculated meansigma methods in angle is as the vertical dip angle of car plate;
VLP correction unit, for according to described level inclination and described vertical dip angle, utilizes affine transformation to described car
Board carries out angle correct.
Based on any of the above-described embodiment, this device also includes: training module, for instructing described character machining device
Practice;Specifically include:
Nominal data unit, for obtaining the character data of various different angles as nominal data;
Training sample unit, for obtaining multiple character image datas of various different visual angles as training sample;
Training unit, for according to described nominal data, utilizes CNN convolutional neural networks model to enter described training sample
Row training, obtains character machining device.
Wherein, based on each embodiment above-mentioned, this device can be integrated in intellectual analysis one ball machine, or embeds intelligence
In energy analytical equipment, embedded intelligence analytical equipment can be connected formation car plate detecting system by switch with common ball machine.
In description, each embodiment uses the mode gone forward one by one to describe, and what each embodiment stressed is real with other
Executing the difference of example, between each embodiment, identical similar portion sees mutually.For device disclosed in embodiment
Speech, owing to it corresponds to the method disclosed in Example, so describe is fairly simple, relevant part sees method part explanation
?.
Professional further appreciates that, in conjunction with the unit of each example that the embodiments described herein describes
And algorithm steps, it is possible to electronic hardware, computer software or the two be implemented in combination in, in order to clearly demonstrate hardware and
The interchangeability of software, the most generally describes composition and the step of each example according to function.These
Function performs with hardware or software mode actually, depends on application-specific and the design constraint of technical scheme.Specialty
Technical staff specifically should can be used for using different methods to realize described function to each, but this realization should not
Think beyond the scope of this invention.
The method described in conjunction with the embodiments described herein or the step of algorithm can direct hardware, processor be held
The software module of row, or the combination of the two implements.Software module can be placed in random access memory (RAM), internal memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, depositor, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
The method and device detected car plate provided by the present invention above is described in detail.Tool used herein
Principle and the embodiment of the present invention are set forth by body example, and the explanation of above example is only intended to help and understands this
Bright method and core concept thereof.It should be pointed out that, for those skilled in the art, without departing from the present invention
On the premise of principle, it is also possible to the present invention is carried out some improvement and modification, these improve and modification also falls into right of the present invention
In the protection domain required.
Claims (10)
1. the method for a car plate detection, it is characterised in that including:
Obtain the image information that image capture device gathers;
Utilize character machining device that described image information is detected, obtain the zone position of each character in described image information
Breath and the character kind of information of correspondence;
Zone position information according to each described character and the character kind of information of correspondence and car plate rule are to described image information
In car plate be fitted obtaining matching car plate, and described matching car plate carried out License Plate obtain License Plate result;
According to described License Plate result, described car plate is carried out angle correct;
Described car plate after correction is carried out character recognition, obtains described license board information.
2. the method for car plate detection as claimed in claim 1, it is characterised in that according to the zone position information of each described character
And the character kind of information of correspondence and car plate rule are fitted obtaining matching car plate to the car plate in described image information, bag
Include:
Zone position information according to each described character and the character kind of information of correspondence, determine the described of chinese character and correspondence
The zone position information of chinese character;
Whether alphabetic character is there is in the range of the zone position information preset distance of chinese character described in judging distance;
If existing, then determine according to the zone position information of the zone position information of described chinese character and the alphabetic character of correspondence
Predetermined car plate direction;
Described predetermined car plate direction determines the zone position information of the alphabetic character of remaining five alphabetic character and correspondence,
Matching car plate to seven characters meeting car plate rule.
3. the method for car plate detection as claimed in claim 2, it is characterised in that described matching car plate is carried out License Plate and obtains
To License Plate result, including:
According to the zone position information of each character in described matching car plate, ask for the boundary rectangle of described matching car plate;
Described boundary rectangle is enlarged according to predetermined ratio, obtains License Plate result.
4. the method for car plate detection as claimed in claim 3, it is characterised in that according to described License Plate result to described car
Board carries out angle correct, including:
Determine the central point of the zone position information of each character in described matching car plate, fit to first according to each described central point
Straight line, and the level inclination of car plate is determined according to described first straight line and horizontal angle;
Calculate the vertical dip angle of each character in described matching car plate, and will be according to the calculated meansigma methods conduct of each vertical dip angle
The vertical dip angle of car plate;
According to described level inclination and described vertical dip angle, utilize affine transformation that described car plate is carried out angle correct.
5. the method for the car plate detection as described in any one of Claims 1-4, it is characterised in that the instruction of described character machining device
Practice method, including:
Obtain the character data of various different angles as nominal data;
Obtain multiple character image datas of various different visual angles as training sample;
According to described nominal data, utilize CNN convolutional neural networks model that described training sample is trained, obtain character inspection
Survey device.
6. the device of a car plate detection, it is characterised in that including:
Acquisition module, for obtaining the image information that image capture device gathers;
Character machining module, is used for utilizing character machining device to detect described image information, obtains in described image information
The zone position information of each character and the character kind of information of correspondence;
Car plate matching locating module, for according to the zone position information of each described character and the character kind of information of correspondence and car
Board rule is fitted obtaining matching car plate to the car plate in described image information, and described matching car plate is carried out License Plate
Obtain License Plate result;
VLP correction module, for carrying out angle correct according to described License Plate result to described car plate;
Recognition of License Plate Characters module, for the described car plate after correction is carried out character recognition, obtains described license board information.
7. the device of car plate detection as claimed in claim 6, it is characterised in that described car plate matching locating module includes car plate
Fitting unit, wherein, described car plate fitting unit includes:
Chinese character subelement, for the zone position information according to each described character and the character kind of information of correspondence, determines
The zone position information of the described chinese character of chinese character and correspondence;
First alphabetic character judgment sub-unit, for the model of the zone position information preset distance of chinese character described in judging distance
Whether alphabetic character is there is in enclosing;
Car plate direction determines subelement, if for existing, then according to zone position information and the word of correspondence of described chinese character
The zone position information of alphabetic character determines predetermined car plate direction;
Car plate matching subelement, for determining remaining five alphabetic characters and alphabetic word of correspondence on described predetermined car plate direction
The zone position information of symbol, obtains meeting the matching car plate of seven characters of car plate rule.
8. the device of car plate detection as claimed in claim 7, it is characterised in that described car plate matching locating module includes car plate
Positioning unit, wherein, described License Plate unit includes:
Boundary rectangle obtains subelement, for according to the zone position information of each character in described matching car plate, asks for described plan
Close the boundary rectangle of car plate;
License Plate subelement, for being enlarged according to predetermined ratio by described boundary rectangle, obtains License Plate result.
9. the device of car plate detection as claimed in claim 8, it is characterised in that described VLP correction module, including:
Level inclination unit, for determining the central point of the zone position information of each character in described matching car plate, according to each institute
State central point and fit to the first straight line, and determine the level inclination of car plate according to described first straight line and horizontal angle;
Vertical dip angle unit, for calculating the vertical dip angle of each character in described matching car plate, and will be according to each vertical dip angle meter
The meansigma methods obtained is as the vertical dip angle of car plate;
VLP correction unit, for according to described level inclination and described vertical dip angle, utilizes affine transformation to enter described car plate
Row angle correct.
10. the device of the car plate detection as described in any one of claim 6 to 9, it is characterised in that also include: training module, use
In described character machining device is trained;Specifically include:
Nominal data unit, for obtaining the character data of various different angles as nominal data;
Training sample unit, for obtaining multiple character image datas of various different visual angles as training sample;
Training unit, for according to described nominal data, utilizes CNN convolutional neural networks model to instruct described training sample
Practice, obtain character machining device.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610556058.3A CN106203418A (en) | 2016-07-14 | 2016-07-14 | A kind of method and device of car plate detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610556058.3A CN106203418A (en) | 2016-07-14 | 2016-07-14 | A kind of method and device of car plate detection |
Publications (1)
Publication Number | Publication Date |
---|---|
CN106203418A true CN106203418A (en) | 2016-12-07 |
Family
ID=57475200
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610556058.3A Pending CN106203418A (en) | 2016-07-14 | 2016-07-14 | A kind of method and device of car plate detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106203418A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106874897A (en) * | 2017-04-06 | 2017-06-20 | 北京精英智通科技股份有限公司 | A kind of licence plate recognition method and device |
CN107220638A (en) * | 2017-07-03 | 2017-09-29 | 深圳市唯特视科技有限公司 | A kind of car plate detection recognition methods based on deep learning convolutional neural networks |
CN108229470A (en) * | 2017-12-22 | 2018-06-29 | 北京市商汤科技开发有限公司 | character image processing method, device, equipment and storage medium |
CN108681729A (en) * | 2018-05-08 | 2018-10-19 | 腾讯科技(深圳)有限公司 | Text image antidote, device, storage medium and equipment |
CN108701234A (en) * | 2018-03-05 | 2018-10-23 | 深圳前海达闼云端智能科技有限公司 | Licence plate recognition method and cloud system |
CN109697439A (en) * | 2017-10-24 | 2019-04-30 | 北京京东尚科信息技术有限公司 | Chinese character localization method and device, storage medium and electronic equipment |
CN109887325A (en) * | 2019-03-19 | 2019-06-14 | 南京维智感网络科技有限公司 | Road-surface concrete field managing device and its method based on Xian Zou mechanism |
CN110059683A (en) * | 2019-04-15 | 2019-07-26 | 广州广电银通金融电子科技有限公司 | A kind of license plate sloped antidote of wide-angle based on end-to-end neural network |
CN110443244A (en) * | 2019-08-12 | 2019-11-12 | 深圳市捷顺科技实业股份有限公司 | A kind of method and relevant apparatus of graphics process |
CN110956169A (en) * | 2018-09-27 | 2020-04-03 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and electronic equipment |
CN111178357A (en) * | 2019-12-31 | 2020-05-19 | 松立控股集团股份有限公司 | License plate recognition method, system, device and storage medium |
CN111178359A (en) * | 2018-11-09 | 2020-05-19 | 中移(杭州)信息技术有限公司 | License plate number recognition method, device and equipment and computer storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1085455A2 (en) * | 1999-09-15 | 2001-03-21 | Siemens Corporate Research, Inc. | License plate recognition with an intelligent camera |
CN102096821A (en) * | 2010-04-15 | 2011-06-15 | 西安理工大学 | Number plate identification method under strong interference environment on basis of complex network theory |
CN102750525A (en) * | 2012-06-20 | 2012-10-24 | 太仓博天网络科技有限公司 | License plate recognition system under real-time traffic |
CN102915433A (en) * | 2012-09-13 | 2013-02-06 | 中国科学院自动化研究所 | Character combination-based license plate positioning and identifying method |
CN105488501A (en) * | 2015-11-26 | 2016-04-13 | 南京富士通南大软件技术有限公司 | Method for correcting license plate slant based on rotating projection |
-
2016
- 2016-07-14 CN CN201610556058.3A patent/CN106203418A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1085455A2 (en) * | 1999-09-15 | 2001-03-21 | Siemens Corporate Research, Inc. | License plate recognition with an intelligent camera |
CN102096821A (en) * | 2010-04-15 | 2011-06-15 | 西安理工大学 | Number plate identification method under strong interference environment on basis of complex network theory |
CN102750525A (en) * | 2012-06-20 | 2012-10-24 | 太仓博天网络科技有限公司 | License plate recognition system under real-time traffic |
CN102915433A (en) * | 2012-09-13 | 2013-02-06 | 中国科学院自动化研究所 | Character combination-based license plate positioning and identifying method |
CN105488501A (en) * | 2015-11-26 | 2016-04-13 | 南京富士通南大软件技术有限公司 | Method for correcting license plate slant based on rotating projection |
Non-Patent Citations (2)
Title |
---|
周丽芳等: "《模式识别原理及工程应用》", 30 April 2013, 机械工业出版社 * |
蒋先刚等: "《基于Delphi的数字图像处理工程软件设计》", 31 December 2006, 中国水利水电出版社 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106874897A (en) * | 2017-04-06 | 2017-06-20 | 北京精英智通科技股份有限公司 | A kind of licence plate recognition method and device |
CN107220638A (en) * | 2017-07-03 | 2017-09-29 | 深圳市唯特视科技有限公司 | A kind of car plate detection recognition methods based on deep learning convolutional neural networks |
CN109697439A (en) * | 2017-10-24 | 2019-04-30 | 北京京东尚科信息技术有限公司 | Chinese character localization method and device, storage medium and electronic equipment |
CN108229470A (en) * | 2017-12-22 | 2018-06-29 | 北京市商汤科技开发有限公司 | character image processing method, device, equipment and storage medium |
US11275961B2 (en) | 2017-12-22 | 2022-03-15 | Beijing Sensetime Technology Development Co., Ltd. | Character image processing method and apparatus, device, and storage medium |
WO2019169532A1 (en) * | 2018-03-05 | 2019-09-12 | 深圳前海达闼云端智能科技有限公司 | License plate recognition method and cloud system |
CN108701234A (en) * | 2018-03-05 | 2018-10-23 | 深圳前海达闼云端智能科技有限公司 | Licence plate recognition method and cloud system |
CN108681729A (en) * | 2018-05-08 | 2018-10-19 | 腾讯科技(深圳)有限公司 | Text image antidote, device, storage medium and equipment |
CN110956169A (en) * | 2018-09-27 | 2020-04-03 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and electronic equipment |
CN111178359A (en) * | 2018-11-09 | 2020-05-19 | 中移(杭州)信息技术有限公司 | License plate number recognition method, device and equipment and computer storage medium |
CN109887325A (en) * | 2019-03-19 | 2019-06-14 | 南京维智感网络科技有限公司 | Road-surface concrete field managing device and its method based on Xian Zou mechanism |
CN110059683A (en) * | 2019-04-15 | 2019-07-26 | 广州广电银通金融电子科技有限公司 | A kind of license plate sloped antidote of wide-angle based on end-to-end neural network |
CN110443244A (en) * | 2019-08-12 | 2019-11-12 | 深圳市捷顺科技实业股份有限公司 | A kind of method and relevant apparatus of graphics process |
CN110443244B (en) * | 2019-08-12 | 2023-12-05 | 深圳市捷顺科技实业股份有限公司 | Graphics processing method and related device |
CN111178357A (en) * | 2019-12-31 | 2020-05-19 | 松立控股集团股份有限公司 | License plate recognition method, system, device and storage medium |
CN111178357B (en) * | 2019-12-31 | 2023-09-29 | 松立控股集团股份有限公司 | License plate recognition method, system, device and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106203418A (en) | A kind of method and device of car plate detection | |
CN109685000A (en) | A kind of method for detecting parking stalls and device of view-based access control model | |
CN106599792B (en) | Method for detecting hand driving violation behavior | |
US9087258B2 (en) | Method for counting objects and apparatus using a plurality of sensors | |
CN103473774B (en) | A kind of vehicle positioning method based on pavement image characteristic matching | |
CN102722711B (en) | License plate character sequence identification method and device in license plate type identification | |
CN104268596B (en) | A kind of Car license recognition device and its detection method of license plate and system | |
CN110501018A (en) | A kind of traffic mark board information collecting method for serving high-precision map producing | |
CN107563265A (en) | A kind of high beam detection method and device | |
CN101877174A (en) | Vehicle speed measurement method, supervisory computer and vehicle speed measurement system | |
CN108388641A (en) | A kind of means of transportation based on deep learning ground drawing generating method and system | |
CN106340044A (en) | Camera external parameter automatic calibration method and calibration device | |
CN101900566A (en) | Pixel-based texture-rich clear path detection | |
CN105574542A (en) | Multi-vision feature vehicle detection method based on multi-sensor fusion | |
CN105279512A (en) | Tilt vehicle license plate recognition method and device | |
CN107273816A (en) | Traffic speed limit label detection recognition methods based on vehicle-mounted forward sight monocular camera | |
CN109697441A (en) | A kind of object detection method, device and computer equipment | |
CN104331691A (en) | Vehicle logo classifier training method, vehicle logo recognition method and device | |
CN107316332A (en) | The camera and scene relating scaling method and system of a kind of application intelligent driving | |
CN111295666A (en) | Lane line detection method, device, control equipment and storage medium | |
CN107133623A (en) | A kind of pointer position accurate detecting method positioned based on background subtraction and the center of circle | |
CN110288006A (en) | A kind of license plate number automatic Verification method and system | |
CN108960083A (en) | Based on automatic Pilot objective classification method combined of multi-sensor information and system | |
CN108288020A (en) | Video shelter detecting system based on contextual information and method | |
CN104463238B (en) | A kind of automobile logo identification method and system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161207 |
|
RJ01 | Rejection of invention patent application after publication |