CN104408454A - License plate character segmentation method based on elastic template matching algorithm - Google Patents
License plate character segmentation method based on elastic template matching algorithm Download PDFInfo
- Publication number
- CN104408454A CN104408454A CN201410594829.9A CN201410594829A CN104408454A CN 104408454 A CN104408454 A CN 104408454A CN 201410594829 A CN201410594829 A CN 201410594829A CN 104408454 A CN104408454 A CN 104408454A
- Authority
- CN
- China
- Prior art keywords
- sequence
- car plate
- opt
- matching
- forming board
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
- G06V10/267—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion by performing operations on regions, e.g. growing, shrinking or watersheds
-
- 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
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Character Discrimination (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to a license plate character segmentation method based on an elastic template matching algorithm. An ideal elastic template is described and a dimension change of the license plate in the video can be described well; and thus the character segmentation problem is presented into an optimization problem. A license plate picture is matched with the ideal template and global optimized matching can be obtained from the license plate picture, thereby effectively solving various problems in the traditional license plate character segmentation. Corresponding positions of all characters in the license plate picture are determined based on all prior character positions of the template, so that influences of special character projection, non-uniform illumination or weather, and noises can be eliminated. Moreover, the segmentation algorithm can be applied to the binary projection sequence; the matching speed is fast; and no character identification time is increased. According to the invention, the method with low requirement on the imaging quality is suitable for real-time identification on the license plate character.
Description
Technical field
The invention belongs to digital image processing techniques.
Background technology
Automotive number plate identification is called for short Car license recognition, and Car license recognition refers to and vehicle that pavement of road runs at high speed can be detected and the information (containing Chinese character, English alphabet, arabic numeral and number plate color) automatically extracting vehicle license carries out the technology that processes.License plate recognition technology is modern intelligent transportation system important component part, and its application is very extensive.It is based on the technology such as computer vision process, Digital Image Processing, pattern-recognition, carries out Treatment Analysis, obtain the number-plate number of each car, thus complete identifying the vehicle image of shot by camera or video image.Wherein, to the correct segmentation of characters on license plate be the prerequisite of Recognition of License Plate Characters.
But to split accurately characters on license plate and still there is a following difficult problem:
1, according to " People's Republic of China's automotive number plate " (GA36-2007), containing 7 characters in automotive number plate, and except " 1 ", the width of other characters is roughly the same; In the character that width is roughly the same, character " J ", " T " and " L " are compared with other characters, there is its singularity, such as ask to horizontal direction the individual pixel that do not have of " L ", then sum peaks will concentrate on a part of erecting in left side corresponding " L ", and a horizontal corresponding part generally can be taken as noise when carrying out binaryzation and be removed, character " 0 " is if adaptive threshold is too high for another example, " 0 " may be regarded as two " 1 " after binaryzation; The disconnected character of minority is there is, as " river " etc. in the character of automotive number plate.
2, the external environment residing for Vehicle License Plate Recognition System is complicated and changeable.Vehicle License Plate Recognition System under the inclement weather conditions such as rainy day, dense fog, heavy snow, due to atrocious weather car plate can be caused to be blocked or vehicle Facing Movement time the car light car plate uneven illumination that causes.
3, in highway and incity, city actual application, the car plate be difficult to involved by ensureing all does not have stained, car plate is after use several years, there will be the phenomenons such as pollution and wearing and tearing unavoidably, and the vehicle travelled on road surface to be also difficult to guarantee be all the clean car plate of standard, therefore in actual environment, in the face of damaged dirty old car plate, the recognition capability how improving Vehicle License Plate Recognition System is also the problem that actual needs solves.
The situations such as these factors affect the quality obtaining license plate image all greatly above, cause actual acquisition character to there is adhesion, stained.Run into the problems referred to above, traditional partitioning algorithm or judge split and unsuccessfully and not return any information, or carried out wrong segmentation because make use of error message locally and returned full of prunes information.
Summary of the invention
Technical scheme to be solved by this invention is, provide a kind of based on car plate picture with based on automotive number plate standard and the global optimum that can describe the desirable template of car plate dimensional variation in video mate, thus complete the method for License Plate Character Segmentation.
The present invention is based on the registration number character dividing method of Elastic forming board matching algorithm, it is characterized in that, comprise following steps for solving the problems of the technologies described above adopted technical scheme:
Step 1, by moving object segmentation, obtain the license plate image of car plate position of coarse positioning;
The often row summation of the license plate image of step 2, right coarse positioning, by by the row of each row with compare to come binary conversion treatment with adaptive threshold and obtain binaryzation car plate sequence, in binaryzation car plate sequence, 0 represents background, and 1 represents characters on license plate;
Step 3, binaryzation car plate sequence to be mated with the Elastic forming board meeting car plate standard, thus realize the rectification to binaryzation car plate sequence;
Matching process is completed by following interative computation:
Wherein, i represents i-th of binaryzation car plate sequence, i=1,2 ..., n, n is the length of binaryzation car plate sequence, and j represents the jth position of sequence in Elastic forming board, j=1,2 ..., m, m is the length of sequence in Elastic forming board, and OPT (i, j) represents the Optimum Matching of the i-th bit element to the jth bit element of sequence in Elastic forming board of binaryzation car plate sequence, min represents and minimizes, and δ is room penalty value
for coupling penalty value, when the i-th bit element of binaryzation car plate sequence is identical with Elastic forming board jth position, when namely correctly mating
be 0, when the i-th bit element of binaryzation car plate sequence is different from Elastic forming board jth position,
for erroneous matching penalty value
; Initial value OPT (0,0)=0, OPT (i, 0)=i δ, OPT (0, j)=j δ is set;
As the n-th bit calculating binaryzation car plate sequence plain Optimum Matching OPT (n, m) to the m bit element of sequence in Elastic forming board, coupling terminates;
Step 4, by carrying out the Converse solved matching status obtaining iteration each time to Optimum Matching OPT (n, m), described matching status is:
When
then represent that the i-th bit element of current binaryzation car plate sequence is correct coupling to the matching status of the jth bit element of sequence in Elastic forming board;
When
then represent that the i-th bit element of current binaryzation car plate sequence is erroneous matching to the matching status of the jth bit element of sequence in Elastic forming board;
When
then represent plug hole on the i-th bit element of current binaryzation car plate sequence;
When
plug hole on the jth bit element of expression sequence in current Elastic forming board;
By the Converse solved shape obtaining Elastic forming board sequence under Optimum Matching, thus determined each character position in the input car plate sequence of Corresponding matching by the Elastic forming board sequence determining shape, finally realize the segmentation to characters on license plate.
The invention describes a desirable Elastic forming board, meet China's automotive number plate standard, by achieving the elasticity adjustment of the shape of template sequence to plug hole in the sequence continuing to mate, well can describe car plate dimensional variation in video, Character segmentation problem is expressed as optimization problem.Car plate picture is mated with desirable template, from car plate picture, obtains global optimum mate the various problems can effectively tackled in traditional License Plate Character Segmentation.The correspondence position of each character in car plate picture is determined by each character position of priori in template, not by the projection of the character of singularity, the noise effect of uneven illumination or weather.And partitioning algorithm can act on the projection sequence of binaryzation, matching speed is exceedingly fast, and does not increase the time of character recognition.
The invention has the beneficial effects as follows, split accurate rate higher, travelling speed is fast, requires lower, be applicable to the Real time identification of characters on license plate to image quality.
Accompanying drawing explanation
Fig. 1 is embodiment schematic flow sheet.
Embodiment
For convenience of describing embodiment content, first some terms are described here:
Meet the Elastic forming board of car plate standard: China 07 formula characters on license plate region is made up of 7 characters and 1 blank character.The standard outer profile size of automobile single file licence plate: long 440mm and wide 140mm; Characters on license plate total length is 409mm, and single character uniform width is 45mm, height is 90mm; Interval 34mm between second and three-character doctrine, all the other character pitches are 12mm; The wide 10mm round dot of blank character, second and the spacing of three-character doctrine and blank character be 12mm; Stroke width 10mm.The plate template of series as standard of the two-value that a use " 0,1 " represents is generated according to the proportionate relationship of character each in car plate standard and character pitch.Because in reality, car plate has dimensional variation in scene, therefore this template must be the proportionate relationship of description standard car plate, has again certain elasticity.
Moving object segmentation technology: have multi-motion object detection technology, general is all utilize continuous print multiframe in video flowing, modeling is carried out to prospect and moving object and background and stationary object, then in conjunction with level and smooth partitioning algorithm, the pixel in present frame is split.Current many employing mixed Gaussian background modeling technology.
Binaryzation car plate sequence: sued for peace to a specific direction by image, then adopts the binary conversion treatment of adaptive threshold, namely obtains binaryzation car plate sequence to summation image.This sequence saves enough carve informations on the one hand, the segmentation problem simultaneously will be converted into the segmentation problem of two dimensional image one dimension binaryzation sequence again.Which greatly enhances processing speed.
Character classifier: the present invention adopts support vector machines when character recognition.SVM theory provides a kind of complicacy avoiding higher dimensional space, directly with Product function in this space (being kernel function), recycles the method for solving when linear separability, the decision problem of the higher dimensional space that direct solution is corresponding.The main thought of SVM may be summarized to be: it is that linear can a point situation be analyzed, for the situation of linearly inseparable, by using non-linear map the sample of low-dimensional input space linearly inseparable is converted into high-dimensional feature space makes its linear separability, thus make high-dimensional feature space adopt linear algorithm to carry out linear analysis to the nonlinear characteristic of sample to become possibility.
Based on the Car license recognition of Elastic forming board matching algorithm, as shown in Figure 1, following step is comprised:
Step 1, car plate coarse positioning
Step 1-1: moving object segmentation.First Vehicle License Plate Recognition System receives the image data stream that monitoring equipment collects.Due to the moving object that automobile is larger in scene, so first by the most contiguous for image down-sampling, the speed that less picture significantly can improve coarse positioning can be processed.Then the image of down-sampling is carried out background modeling, therefrom detect moving object.
Step 1-2: obtain car plate rough position.Each block moving object region is processed separately.Inherent characteristic based on car plate: license plate area has the saltus step of relative abundance, but not license plate area relative smooth.Utilize the characteristic that car plate marginal information is abundant, use edge detecting technology to process the image obtained, combining form closed operation and algorithm of convex hull obtain the rough position of car plate.Then affined transformation adjustment car plate attitude is carried out.This step is to the thick extraction of car plate position, and the existing method that can extract car plate position is all applicable.
Step 2, License Plate Character Segmentation
Obtain the car plate position of single unit vehicle in the target area of monitoring after, the accurate segmentation of characters on license plate can be realized as follows.
Step 2-1: first sue for peace to car plate Width to the single license plate image of the coarse positioning obtained in step 1, then utilizes adaptive threshold to carry out binary conversion treatment, will obtain car plate sequence like this, and wherein 0 represents background, and 1 represents characters on license plate.According to existing car plate standard, a template sequence can be set, again because car plate size in scene has certain change, so this template must have certain elasticity meeting car plate standard.
Step 2-2: the template sequence of setting is mated with the car plate sequence of binaryzation.Thus the rectification realized binaryzation car plate sequence.Then just split position accurately can be obtained by rectification sequence.Specific algorithm is as follows:
Definition X (x
1, x
2..., x
n) represent input car plate sequence, Y (y
1, y
2..., y
m) representing Elastic forming board sequence, defconstant δ represents the penalty value in a room, defconstant
represent the penalty value of erroneous matching one, then
for coupling penalty value, work as x
i=y
jtime,
work as x
i≠ y
jtime,
i represents i-th of binaryzation car plate sequence, i=1,2 ..., n, n are the length of binaryzation car plate sequence, and j represents the jth position of sequence in Elastic forming board, j=1,2 ..., m, m are the length of sequence in Elastic forming board.Definition finally defines OPT (i, j) and represents that the i-th bit element of binaryzation car plate sequence is to the Optimum Matching of the jth bit element of sequence in Elastic forming board, and min represents and minimizes.
Matching process is completed by following interative computation:
Min represents and minimizes, and arranges initial value OPT (0,0)=0, OPT (i, 0)=i δ, OPT (0, j)=j δ; As the n-th bit calculating binaryzation car plate sequence plain Optimum Matching OPT (n, m) to the m bit element of sequence in Elastic forming board, coupling terminates;
If by OPT (i, j) i=1,2,, n, j=1,2 ..., the computation process of m is considered as filling in a form, each element in form is OPT (i, j) value, and each element coordinate is in the table exactly (i, j), shown in formula, the value that will calculate (i, j) place to one must first calculate (i-1, j) in form, (i, j-1), (i-1, j-1) three value at place, so entirety is filled in a form, the order of lattice is: from left to right, from top to bottom; Namely OPT (1,1) is first calculated, OPT (1,2) ... OPT (1, m), then calculate OPT (2,1) ..., OPT (2, m).
For each in car plate sequence and template, in last matching result, only have 3 class situations: 1. have contraposition, and to the same x of bit sign
i=y
j2. there is contraposition, but to the different x of bit sign
i≠ y
j, 3. pair room.
Such as:
Input car plate sequence 1010011, and template sequence is 00110011, so matching result is:
10 10011
00110011
1st is situation 2 erroneous matching, and the 3rd is situation 3 pairs of rooms (on input car plate sequence the 3rd bit element plug holes), and other position is all correct coupling.
For another example:
Input car plate sequence 111111110011100011100111001110011100111001111111111, and template sequence
Be 00111001110001110011100111001110011100, so matching result is:
111111111111100011100101000100011100111001111111111
00111001110001110011100111001110011100
Suppose that input car plate is: river A B0123.Can see in list entries there is Railway Project.
1. general input car plate sequence can be longer than masterplate, and sequence two ends have from the continuous print 0 or 1 outside car plate.2.0 to be treated to be 101,1 be treated to 010, is not desirable 111.3. river A two Characters Stuck, may because illumination or the reason such as stained.
Described elasticity can be out of shape in other words, namely room, is on the other hand and can not the meaning of gross distortion.In the matching process in order to simulate the elasticity of maintenance template sequence and don't make template sequence broken, need to increase some to constant δ to control, after the matching status of continuous k iteration is plug hole, increase room penalty value δ or room penalty value δ is adjusted to infinitely great, afterwards, when the non-plug hole of matching status, be predetermined constant by δ readjustment.
Step 2-3: by carrying out the Converse solved matching status obtaining iteration each time to Optimum Matching OPT (n, m), described matching status is:
When
then represent that the i-th bit element of current binaryzation car plate sequence is correct coupling to the matching status of the jth bit element of sequence in Elastic forming board;
When
then represent that the i-th bit element of current binaryzation car plate sequence is erroneous matching to the matching status of the jth bit element of sequence in Elastic forming board;
When
then represent plug hole on the i-th bit element of current binaryzation car plate sequence;
When
plug hole on the jth bit element of expression sequence in current Elastic forming board;
By the Converse solved shape obtaining Elastic forming board sequence under Optimum Matching, thus determined each character position in the input car plate sequence of Corresponding matching by the Elastic forming board sequence determining shape, finally realize the segmentation to characters on license plate.
One paths of Converse solved from OPT (n, m) to OPT (0,0), this paths has just pointed in the direction of each OPT (i, j), the selection of min.Selected by this, just can construct matching sequence.
Step 3, Recognition of License Plate Characters
Step 3-1: utilize the split position obtained to take out character from the former figure of car plate, character is sent in sorter and carry out character recognition.First known character is normalized, then extracts character feature, send into the sorter built and train, obtain the training result of each character.
Step 3-2: be normalized by the separating character obtained from former figure, then extracts the feature of characters on license plate, uses the sorter trained to carry out Recognition of License Plate Characters.
Said method is carried out on Matlab directly perceived, effective proof of algorithm, be transplanted to afterwards (Visual Studio2010) on C++ platform, write Car license recognition software, carry out actual test.Through the test of mass efficient, compared with traditional algorithm, well solve the character that has specific characteristics in car plate and because the problem of character focal adhesion in environment reason car plate picture, and travelling speed can process soon in real time, has very high practical value.
Claims (2)
1. based on the registration number character dividing method of Elastic forming board matching algorithm, it is characterized in that, comprise following steps:
Step 1, by moving object segmentation, obtain the license plate image of car plate position of coarse positioning;
The often row summation of the license plate image of step 2, right coarse positioning, by by the row of each row with compare to come binary conversion treatment with adaptive threshold and obtain binaryzation car plate sequence, in binaryzation car plate sequence, 0 represents background, and 1 represents characters on license plate;
Step 3, binaryzation car plate sequence to be mated with the Elastic forming board meeting car plate standard, thus realize the rectification to binaryzation car plate sequence;
Matching process is completed by following interative computation:
Wherein, i represents i-th of binaryzation car plate sequence, i=1,2 ..., n, n are the length of binaryzation car plate sequence, and j represents the jth position of sequence in Elastic forming board, j=1,2 ..., m, m are the length of sequence in Elastic forming board; OPT (i, j) represents that the i-th bit element of binaryzation car plate sequence is to the Optimum Matching of the jth bit element of sequence in Elastic forming board, and min represents and minimizes, and δ is room penalty value,
for coupling penalty value, when the i-th bit element of binaryzation car plate sequence is identical with Elastic forming board jth position, when namely correctly mating
be 0, when the i-th bit element of binaryzation car plate sequence is different from Elastic forming board jth position,
for erroneous matching penalty value
initial value OPT (0,0)=0, OPT (i, 0)=i δ, OPT (0, j)=j δ is set;
As the n-th bit calculating binaryzation car plate sequence plain Optimum Matching OPT (n, m) to the m bit element of sequence in Elastic forming board, coupling terminates;
Step 4, by carrying out the Converse solved matching status obtaining iteration each time to Optimum Matching OPT (n, m), described matching status is:
When
then represent that the i-th bit element of current binaryzation car plate sequence is correct coupling to the matching status of the jth bit element of sequence in Elastic forming board;
When
then represent that the i-th bit element of current binaryzation car plate sequence is erroneous matching to the matching status of the jth bit element of sequence in Elastic forming board;
As OPT (i, j)=δ+OPT (i-1, j), then represent plug hole on the i-th bit element of current binaryzation car plate sequence;
As OPT (i, j)=δ+OPT (i, j-1), plug hole on the jth bit element representing sequence in current Elastic forming board;
By the Converse solved shape obtaining Elastic forming board sequence under Optimum Matching, thus determined each character position in the input car plate sequence of Corresponding matching by the Elastic forming board sequence determining shape, finally realize the segmentation to characters on license plate.
2. as claimed in claim 1 based on the registration number character dividing method of Elastic forming board matching algorithm, it is characterized in that, in step 3 matching process, after the matching status of continuous k iteration is plug hole, increase room penalty value δ, afterwards, when the non-plug hole of matching status, be predetermined constant by δ readjustment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410594829.9A CN104408454B (en) | 2014-06-30 | 2014-10-29 | Registration number character dividing method based on Elastic forming board matching algorithm |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410306229 | 2014-06-30 | ||
| CN2014103062298 | 2014-06-30 | ||
| CN201410594829.9A CN104408454B (en) | 2014-06-30 | 2014-10-29 | Registration number character dividing method based on Elastic forming board matching algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104408454A true CN104408454A (en) | 2015-03-11 |
| CN104408454B CN104408454B (en) | 2017-10-17 |
Family
ID=52646085
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410594829.9A Active CN104408454B (en) | 2014-06-30 | 2014-10-29 | Registration number character dividing method based on Elastic forming board matching algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104408454B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488508A (en) * | 2015-12-01 | 2016-04-13 | 天津普达软件技术有限公司 | Method for intelligently segmenting film characters of soft pharmaceutical bag |
| CN105631449A (en) * | 2015-12-21 | 2016-06-01 | 华为技术有限公司 | Method, device and equipment for segmenting picture |
| CN109598271A (en) * | 2018-12-10 | 2019-04-09 | 北京奇艺世纪科技有限公司 | A kind of character segmentation method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101373519A (en) * | 2007-08-20 | 2009-02-25 | 富士通株式会社 | Device and method for recognizing character |
| US20120275653A1 (en) * | 2011-04-28 | 2012-11-01 | Industrial Technology Research Institute | Method for recognizing license plate image, and related computer program product, computer-readable recording medium, and image recognizing apparatus using the same |
| CN103198315A (en) * | 2013-04-17 | 2013-07-10 | 南京理工大学 | License plate character segmentation algorithm based on character outline and template matching |
| CN103207998A (en) * | 2012-12-24 | 2013-07-17 | 电子科技大学 | License plate character segmentation method based on support vector machine |
-
2014
- 2014-10-29 CN CN201410594829.9A patent/CN104408454B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101373519A (en) * | 2007-08-20 | 2009-02-25 | 富士通株式会社 | Device and method for recognizing character |
| US20120275653A1 (en) * | 2011-04-28 | 2012-11-01 | Industrial Technology Research Institute | Method for recognizing license plate image, and related computer program product, computer-readable recording medium, and image recognizing apparatus using the same |
| CN103207998A (en) * | 2012-12-24 | 2013-07-17 | 电子科技大学 | License plate character segmentation method based on support vector machine |
| CN103198315A (en) * | 2013-04-17 | 2013-07-10 | 南京理工大学 | License plate character segmentation algorithm based on character outline and template matching |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105488508A (en) * | 2015-12-01 | 2016-04-13 | 天津普达软件技术有限公司 | Method for intelligently segmenting film characters of soft pharmaceutical bag |
| CN105631449A (en) * | 2015-12-21 | 2016-06-01 | 华为技术有限公司 | Method, device and equipment for segmenting picture |
| CN109598271A (en) * | 2018-12-10 | 2019-04-09 | 北京奇艺世纪科技有限公司 | A kind of character segmentation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104408454B (en) | 2017-10-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105260699B (en) | A kind of processing method and processing device of lane line data | |
| CN102663377B (en) | Character recognition method based on template matching | |
| CN105608456B (en) | A kind of multi-direction Method for text detection based on full convolutional network | |
| CN107967695B (en) | A kind of moving target detecting method based on depth light stream and morphological method | |
| CN102629322B (en) | Character feature extraction method based on stroke shape of boundary point and application thereof | |
| CN105335702B (en) | A kind of bayonet model recognizing method based on statistical learning | |
| CN102509091B (en) | Airplane tail number recognition method | |
| CN104715252B (en) | A kind of registration number character dividing method of dynamic template combination pixel | |
| CN110619327A (en) | Real-time license plate recognition method based on deep learning in complex scene | |
| CN103268489A (en) | Motor vehicle plate identification method based on sliding window searching | |
| CN105427626A (en) | Vehicle flow statistics method based on video analysis | |
| CN105354568A (en) | Convolutional neural network based vehicle logo identification method | |
| CN105260712A (en) | Method and system for detecting pedestrian in front of vehicle | |
| CN109341580A (en) | Method for tracing, system and the device of profile of steel rail | |
| CN110516633B (en) | Lane line detection method and system based on deep learning | |
| CN107491720A (en) | A kind of model recognizing method based on modified convolutional neural networks | |
| CN104268506A (en) | Passenger flow counting detection method based on depth images | |
| CN103065138A (en) | Recognition method of license plate number of motor vehicle | |
| CN101673338A (en) | Fuzzy license plate identification method based on multi-angle projection | |
| CN103810505A (en) | Vehicle identification method and system based on multilayer descriptors | |
| CN104200228A (en) | Recognizing method and system for safety belt | |
| CN105488811A (en) | Depth gradient-based target tracking method and system | |
| CN104299009A (en) | Plate number character recognition method based on multi-feature fusion | |
| CN104200207A (en) | License plate recognition method based on Hidden Markov models | |
| CN101727654A (en) | Method realized by parallel pipeline for performing real-time marking and identification on connected domains of point targets |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210518 Address after: 610041 China (Sichuan) pilot Free Trade Zone No.3, 11th floor, building 6, no.599, shijicheng South Road, Chengdu high tech Zone Patentee after: Houpu clean energy Co.,Ltd. Address before: 611731, No. 2006, West Avenue, Chengdu hi tech Zone (West District, Sichuan) Patentee before: University of Electronic Science and Technology of China |
|
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: No.3, 11th floor, building 6, no.599, shijicheng South Road, Chengdu hi tech Zone, China (Sichuan) pilot Free Trade Zone, Chengdu, Sichuan 610041 Patentee after: Houpu clean energy (Group) Co.,Ltd. Address before: 610041 China (Sichuan) pilot Free Trade Zone No.3, 11th floor, building 6, no.599, shijicheng South Road, Chengdu high tech Zone Patentee before: Houpu clean energy Co.,Ltd. |