CN101630360B - Method for identifying license plate in high-definition image - Google Patents
Method for identifying license plate in high-definition image Download PDFInfo
- Publication number
- CN101630360B CN101630360B CN 200810040548 CN200810040548A CN101630360B CN 101630360 B CN101630360 B CN 101630360B CN 200810040548 CN200810040548 CN 200810040548 CN 200810040548 A CN200810040548 A CN 200810040548A CN 101630360 B CN101630360 B CN 101630360B
- Authority
- CN
- China
- Prior art keywords
- image
- pixel
- character
- license plate
- value
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Character Input (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a method for identifying a license plate in a high-definition image, which comprises steps of candidate license plate area detection, grayscale image binarization, obliquity correction, character segmentation, character identification and post-processing. Compared with the prior art, the method for identifying the license plate in the high-definition image has the advantages of simultaneous and quick positioning and identification of a plurality of license plates in a complex background with high definition (more than 5,000,000 pixels), good image binarization effect under the condition of non-uniform illumination, variability of license plate obliquity and size in a relatively larger range, and the like, and simultaneously greatly reduces the requirements on the installation of a video camera.
Description
Technical field
The present invention relates to car plate identification, particularly relate to a kind of method of in HD image, discerning car plate.
Background technology
The technical grade high-definition camera of hundred everything elements above (1,300,000~5,000,000) begins to be used for the candid photograph to vehicle image in recent years, because it can provide law enforcement required strong evidence, therefore begins to be applied in security protection and intelligent transportation field.
Existing SD image (below the 45 everything elements) licence plate recognition method is difficult to the car plate identification to high-definition image.Reason is that the car plate identification of high-definition image is very high owing to its resolution; Viewfinder range is wide, background is complicated; The car plate positioning difficulty is big, have a plurality of car plates to need identification simultaneously in the image, and single car plate in the single background is generally only discerned in existing SD image car plate identification.In addition, practical licence plate recognition method not only requires discrimination high, and recognition speed wants fast.Therefore how discerning a plurality of number-plate numbers in the high-definition image rapidly and accurately simultaneously, then is that problem to be solved is arranged in the prior art.
Summary of the invention
Technical matters to be solved by this invention is exactly for the defective that overcomes above-mentioned prior art existence a kind of method of in HD image, discerning car plate to be provided.
The object of the invention can be realized through following technical scheme: a kind of method of in HD image, discerning car plate, it is characterized in that, and may further comprise the steps:
(1) with certain zoom factor; The original image captured of convergent-divergent video camera repeatedly, every convergent-divergent image once detects the image of a license plate area in scaled images; With detected license plate area image mapped to original image; Through filtering and merge the coincidence pattern picture, according to the characteristic strength of each license plate area image each image is sorted, obtain all license plate area images;
(2) be gray level image with all license plate area image transitions, carry out binary conversion treatment again, obtain binary image;
(3) degree of tilt of correction binary image;
(4) cut apart character in the binary image;
(5) discern the character that is partitioned into, and check.
Described zoom factor comprises horizontal scaling coefficient and vertically scale coefficient.
Described horizontal scaling coefficient and vertically scale coefficient are 0.5.
The image of described detection license plate area comprises:
Color notation conversion space;
The rim detection location.
Described step (2) adopts the GLLT algorithm to carry out binary conversion treatment.
Described step (5) adopts the artificial neural network recognizer that character is discerned, and according to the arrangement regulation of car plate, recognition result is checked.
Compared with prior art; The present invention has advantages such as the quick simultaneously location identification of many car plates under high resolving power (reaching more than the 500 everything elements) complex background, effective, the license plate sloped degree of inhomogeneous illumination condition hypograph binaryzation and the big I of licence plate change in a big way, has significantly reduced the requirement of video camera installation simultaneously.
Description of drawings
Fig. 1 is the process flow diagram of the embodiment of the invention;
Fig. 2 is that the image pyramid of the embodiment of the invention decomposes and the fusion synoptic diagram;
Fig. 3 is that pixel concerns the field synoptic diagram in the GLLT algorithm of the embodiment of the invention;
Fig. 4 is the GLLT algorithm of the embodiment of the invention and the comparison diagram as a result of histogram binaryzation algorithm;
Fig. 5 is the degree of tilt correction principle figure one of the embodiment of the invention;
Fig. 6 is the degree of tilt correction principle figure two of the embodiment of the invention;
Fig. 7 is a synoptic diagram before and after the degree of tilt of the embodiment of the invention is proofreaied and correct.
Embodiment
Below in conjunction with accompanying drawing the present invention is described further
Shown in Fig. 1~7, a kind of method of in HD image, discerning car plate may further comprise the steps:
(1) with certain zoom factor; The original image captured of convergent-divergent video camera repeatedly, every convergent-divergent image once detects the image of a license plate area in scaled images; With detected license plate area image mapped to original image; Through filtering and merge the coincidence pattern picture, according to the characteristic strength of each license plate area image each image is sorted, obtain all license plate area images;
(2) be gray level image with all license plate area image transitions, carry out binary conversion treatment again, obtain binary image;
(3) degree of tilt of correction binary image;
(4) cut apart character in the binary image;
(5) discern the character that is partitioned into, and check.
Described zoom factor comprises horizontal scaling coefficient and vertically scale coefficient; Described horizontal scaling coefficient and vertically scale coefficient are 0.5; The image of described detection license plate area comprises:
Color notation conversion space; The rim detection location.
Described step (2) adopts the GLLT algorithm to carry out binary conversion treatment; Described step (5) adopts the artificial neural network recognizer that character is discerned, and according to the arrangement regulation of car plate, recognition result is checked.
Embodiment
One, candidate's license plate area detects
In car plate recognition application system, the position that video camera is installed and the scene size of collection have determined the size of car plate in image.When video camera was installed in the side, the car plate size also can differ greatly in the same sub-picture.If it is too strict that the car plate recognizer requires the car plate size, will limit the range of application of car plate identification, bring very burden for the construction debugging.The characters on license plate height setting between 15~50 pixels, can be satisfied most actual requirement of engineering.
In order to detect large scale and undersized car plate simultaneously, adopt the image pyramid formula to decompose, in each grade image, detect the car plate of certain size range.At last pyramid diagrams at different levels are mapped in the original image as detected candidate's license plate area, overlap the zone, sorted according to characteristic strength in all candidate regions, finally obtain all candidate regions through filtering and merging, as shown in Figure 2.The scale-up factor of setting pyramid decomposition level direction and vertical direction is respectively γ
x(<1.0) and γ
y(<1.0),, the first order is decomposed by original image I horizontal direction and vertical direction difference convergent-divergent γ
xAnd γ
yDoubly, obtain first order pyramid image I
1, again by I
1Horizontal direction and vertical direction be convergent-divergent γ respectively
xAnd γ
yDoubly, obtain second level pyramid image I
2..., the rest may be inferred can do N (N=1,2,3 ...) and level decomposes.Generally get γ
x=0.5, γ
y=0.5 so that improve the speed of picture breakdown, when identification character is the high car plate of 15~50 pixels, gets N=2.
Each grade pyramid diagram picture is at first carried out color notation conversion space, to find out the purer car plate of those colors apace.The YUV color space has effect preferably to color separated, the car plate that color is purer, after the conversion brightness stronger, have obvious characteristics.For car plate color not obvious (as in shade), then adopt the vertical edge characteristic detection method, the speed that is characterized in is fast and quite good detecting effectiveness arranged.The YUV color space disturbs less, and car plate is located more quick and precisely, and the method that therefore adopts color space to combine with rim detection is carried out the color space location earlier, carries out the rim detection location again, to realize the location quick and precisely of high-definition image car plate.
Two, gray level image binaryzation
Because it is simpler that gray level image carries out binaryzation than coloured image, generally earlier convert the car plate coloured image that cuts out into gray level image binaryzation again.Binarization method is very many, but does not have method in common.Have only to the characteristics of image in the practical application itself and just can develop the better binary conversion method.In the car plate recognition application,, then can adopt the histogram method binaryzation if illumination is even and contrast is stronger.Histogram method is easy to calculate, and speed is fast.But when uneven illumination is spared, then can't directly use histogram method.When picture contrast is low in addition,, also be difficult to confirm the binaryzation threshold value even illumination is even.
GLLT (the Gray Logical Level Technique) algorithm that the present invention adopts stroke to cut apart to combine with gray-scale statistical.The GLLT algorithm is considered the stroke characteristics of character fully, also considers the gray-scale statistical characteristics of interior character of subrange and background simultaneously, can solve the binaryzation problem of the even low contrast license plate image of uneven illumination effectively.
In the GLLT algorithm; Stroke characteristics according to character; At first divide out with differentiating in the image for pixel that belongs to character stroke and the pixel that belongs to background, the gray-scale value of adding up stroke pixel and background pixels again decides the attribute of the pixel of not divided in the image.The key step of GLLT algorithm is following:
1) (x is that ((x y) is its value after level and smooth to g to picture element in the image for x, gray-scale value y) y) to establish f.Estimate stroke width W (generally getting W=3) according to character boundary in the image, with in the image every be average gray in center calculation (2W+1) * (2W+1) window:
2) establish that (x, y) 8 of W pixel is P in abutting connection with pixel apart from pixel
0, P
1..., P
7(as shown in Figure 3).If g (x, y) than it 4 in abutting connection with pixel P
i, P
(i+4) mod8, P
(i+1) mod8, P
(i+5) mod8(i=0,1,2,3) high T gray level, then (x y) is divided into " white pixel " (value 255); If g (x, y) than it 4 in abutting connection with pixel P
i, P
(i+4) mod8, P
(i+1) mod8, P
(i+5) mod8The T gray level is hanged down in (i=0,1,2,3), and then (x y) is divided into " black pixel " (value 128); Otherwise this pixel is labeled as " unfiled pixel " (value 0).Decision rule:
Wherein H (P) is true, if g (x, y)-f (x, y)>T; L (P) is true, if f (x, y)-g (x, y)>T.Pixel P '
iAnd P '
I+1Be respectively pixel P
iAnd P
I+1(i=0,1,2,3) are over against the pixel of (180 ° of directions).
3) subregion is calculated respectively and is denoted as 255 and 128 the pairing average gray image value of pixel G
1And G
2
4) classify to being denoted as 0 unfiled pixel by following rule:
Fig. 4 has shown the comparison diagram as a result of GLLT algorithm and a few quasi-representative gray level image binaryzation algorithms.The GLLT algorithm adapts to by force, speed is fast.In addition, the GLLT algorithm also has a big advantage, and promptly robustness is stronger, does not have the complicated parameter setting, the more doubt problem of threshold value when having avoided such as the histogram binaryzation.
Three, degree of tilt is proofreaied and correct
The quality that degree of tilt is proofreaied and correct is directly connected to the accuracy of Character segmentation and character recognition, and the degree of tilt correction car plate recognizer The whole calculations in the time proportion bigger.Therefore, developing effective, fireballing degree of tilt correcting algorithm is emphasis and difficult point problem in the car plate recognizer.In car plate identification, mainly contain two types of inclinations: horizontal tilt and vertical bank.Following emphasis is described the degree of tilt correction fast algorithm based on vertical runs length statistics that the present invention proposes, and is example (the horizontal tilt degree is proofreaied and correct similar) with the orthogonal rake correction:
1) finds out character area frame coordinate x
0, x
1, y
0, y
1, and calculate its center point coordinate (x
c, y
c), as shown in Figure 5;
2) off-set value that is set in the up-and-down boundary position is D
k, then (x y) is displaced to (x to picture element
s, y
s) determine (as shown in Figure 6) by following formula:
y
s=y;
3) to given D
kThe displacement diagram picture, calculate the quadratic sum of vertical direction black and white run length;
4) establish D
MaxBe up-and-down boundary position maximum possible off-set value, to [D in the interval
Max,+D
Max] in arbitrary integer off-set value D
k, by the above quadratic sum of calculating vertical direction black and white run length, find out maximal value wherein, then its pairing displacement diagram looks like to be the image after degree of tilt is proofreaied and correct.
Above-mentioned degree of tilt correcting algorithm is very effective, and noise resisting ability is strong.Fig. 7 is synoptic diagram before and after proofreading and correct.
Four, Character segmentation
Character segmentation is to orient the coordinate position of each character in the binary image after proofreading and correct through degree of tilt, adopts vertical projection method.This part difficult point is adhesion (comprising the adhesion of frame and character) and damaged Character segmentation, instructs cutting apart of adhesion and damaged character through parameters such as the mean breadth of estimating character, the average height of character, intercharacter mean distances.
Five, character recognition and aftertreatment
Adopt three layers of feedforward artificial neural network to carry out character recognition,, again these sample training are obtained the network weight coefficient through selecting the great amount of samples character.Artificial neural network is the recognizer of a comparative maturity, and it has, and speed is fast, noise resisting ability is strong, particularly damaged character, character size is had effect preferably than the character recognition of small characters, stroke adhesion.But obtain very high character identification rate, need carry out manual evaluation,, train again to improve constantly character identification rate again through the wrong sample data of continuous collection identification to quantity, the quality of each character sample.
Aftertreatment is that recognition result is further checked, according to the licence plate arrangement regulation, character compares with position and recognition result that numeral possibly occur, finds to be inconsistent the back and further handles.The character that commute is obscured " 8 " and " B ", " D " and " O " increase special discrimination module again and carry out secondary discrimination after identification.
Claims (5)
1. the method for an identification car plate in HD image is characterized in that, may further comprise the steps:
(1) with certain zoom factor; The original image captured of convergent-divergent video camera repeatedly, every convergent-divergent image once detects the image of a license plate area in scaled images; With detected license plate area image mapped to original image; Through filtering and merge the coincidence pattern picture, according to the characteristic strength of each license plate area image each image is sorted, obtain all license plate area images;
(2) be gray level image with all license plate area image transitions, carry out binary conversion treatment again, obtain binary image;
(3) degree of tilt of correction binary image;
(4) cut apart character in the binary image;
(5) discern the character that is partitioned into, and check;
Described step (2) adopts the GLLT algorithm to carry out binary conversion treatment;
Stroke characteristics according to character; At first divide out with differentiating in the image for the pixel that belongs to character stroke and the pixel that belongs to background; The gray-scale value of adding up stroke pixel and background pixels again decides the attribute of the pixel of not divided in the image, and the key step of GLLT algorithm is following:
1) establish f (x, y) be picture element in the image (x, gray-scale value y), g (x is its value after level and smooth y), estimates stroke width W according to character boundary in the image, W=3, with in the image every be average gray in center calculation (2W+1) * (2W+1) window:
2) establish that (x, y) 8 of W pixel is P in abutting connection with pixel apart from pixel
0, P
1..., P
7If g (x, y) than it 4 in abutting connection with pixel P
i, P
(i+4) mod8, P
(i+1) mod8, P
(i+5) mod8(i=0,1,2,3) high T gray level, then (x y) is divided into " white pixel ", value 255: if g (x, y) than it 4 in abutting connection with pixel P
i, P
(i+4) mod8, P
(i+1) mod8, P
(i+5) mod8The T gray level is hanged down in (i=0,1,2,3), and then (x y) is divided into " black pixel ", value 128; Otherwise this pixel is labeled as " unfiled pixel ", value 0; Decision rule:
Plain P
i' and P '
I+1Be respectively pixel P
iAnd P
I+1Over against the pixel of 180 ° of directions, i=0 wherein, 1,2,3;
3) subregion is calculated respectively and is denoted as 255 and 128 the pairing average gray image value of pixel G
1And G
2
4) classify to being denoted as 0 unfiled pixel by following rule:
Degree of tilt based on vertical runs length statistics is proofreaied and correct fast algorithm, is example with the orthogonal rake correction, and the horizontal tilt degree is proofreaied and correct similar:
1) finds out character area frame coordinate x
0, x
1, y
0, y
1, and calculate its center point coordinate (x
c, y
c);
2) off-set value that is set in the up-and-down boundary position is D
k, then (x y) is displaced to (x to picture element
s, y
s) determine by following formula:
y
s=y;
3) to given D
kThe displacement diagram picture, calculate the quadratic sum of vertical direction black and white run length;
4) establish D
MaxBe up-and-down boundary position maximum possible off-set value, to [D in the interval
Max,+D
Max] in arbitrary integer off-set value D
k, by the above quadratic sum of calculating vertical direction black and white run length, find out maximal value wherein, then its pairing displacement diagram looks like to be the image after degree of tilt is proofreaied and correct.
2. a kind of method of in HD image, discerning car plate according to claim 1 is characterized in that described zoom factor comprises horizontal scaling coefficient and vertically scale coefficient.
3. a kind of method of in HD image, discerning car plate according to claim 2 is characterized in that described horizontal scaling coefficient and vertically scale coefficient are 0.5.
4. a kind of method of in HD image, discerning car plate according to claim 2 is characterized in that the image of described detection license plate area comprises:
Color notation conversion space;
The rim detection location.
5. according to claim 1 a kind of in HD image the method for identification car plate, it is characterized in that described step (5) adopts the artificial neural network recognizer that character is discerned, and, recognition result checked according to the arrangement regulation of car plate.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200810040548 CN101630360B (en) | 2008-07-14 | 2008-07-14 | Method for identifying license plate in high-definition image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN 200810040548 CN101630360B (en) | 2008-07-14 | 2008-07-14 | Method for identifying license plate in high-definition image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN101630360A CN101630360A (en) | 2010-01-20 |
CN101630360B true CN101630360B (en) | 2012-12-19 |
Family
ID=41575463
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN 200810040548 Expired - Fee Related CN101630360B (en) | 2008-07-14 | 2008-07-14 | Method for identifying license plate in high-definition image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN101630360B (en) |
Families Citing this family (19)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101937508B (en) * | 2010-09-30 | 2012-07-04 | 湖南大学 | License plate localization and identification method based on high-definition image |
CN102254159B (en) * | 2011-07-07 | 2014-07-02 | 清华大学深圳研究生院 | Interpretation method for digital readout instrument |
CN102708356B (en) * | 2012-03-09 | 2014-04-09 | 沈阳工业大学 | Automatic license plate positioning and recognition method based on complex background |
CN102799862A (en) * | 2012-06-29 | 2012-11-28 | 陕西省交通规划设计研究院 | System and method for pedestrian rapid positioning and event detection based on high definition video monitor image |
CN102810203A (en) * | 2012-06-29 | 2012-12-05 | 陕西省交通规划设计研究院 | System and method for detecting unknown article and event based on high-definition video monitoring image |
CN102902951A (en) * | 2012-06-29 | 2013-01-30 | 陕西省交通规划设计研究院 | System and method for vehicle target location and event detection on basis of high-definition video monitoring images |
CN103632548A (en) * | 2012-08-22 | 2014-03-12 | 上海工程技术大学 | License plate recognition control system and application thereof |
CN102980894B (en) * | 2012-11-13 | 2016-07-06 | 上海交通大学 | Steel structure special type weld nondestructive detection system and method |
CN102982335A (en) * | 2012-12-21 | 2013-03-20 | 北京邮电大学 | Intelligent safe multi-license-plate positioning identification method based on cellular neural network |
CN104408932B (en) * | 2014-11-03 | 2016-08-24 | 河海大学常州校区 | A kind of drunk driving vehicle detecting system based on video monitoring |
CN104599236B (en) * | 2014-12-29 | 2017-10-24 | 小米科技有限责任公司 | A kind of method and apparatus of image rectification |
CN106296578B (en) * | 2015-05-29 | 2020-04-28 | 阿里巴巴集团控股有限公司 | Image processing method and device |
CN105512649A (en) * | 2016-01-22 | 2016-04-20 | 大连楼兰科技股份有限公司 | Method for positioning high-definition video real-time number plate based on color space |
CN110390642B (en) * | 2018-04-20 | 2023-12-08 | 西南民族大学 | Method for geometrically correcting woodcut Tibetan image |
CN109325501B (en) * | 2018-08-14 | 2021-12-03 | 王斌 | Guitar backboard image-based material identification method and device and readable storage medium |
CN109919160B (en) * | 2019-03-04 | 2021-03-23 | 深圳先进技术研究院 | Verification code identification method, device, terminal and storage medium |
CN110414522A (en) * | 2019-07-08 | 2019-11-05 | 深兰科技(上海)有限公司 | A kind of character identifying method and device |
CN111027526B (en) * | 2019-10-25 | 2023-06-13 | 江西省云眼大视界科技有限公司 | Method for improving detection and identification efficiency of vehicle target |
CN112329769A (en) * | 2020-10-27 | 2021-02-05 | 广汽本田汽车有限公司 | Vehicle nameplate identification method and device, computer equipment and storage medium |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1424695A (en) * | 2002-12-24 | 2003-06-18 | 上海高德威智能交通系统有限公司 | Application method for badging image in expressway toll system |
CN1448885A (en) * | 2002-03-28 | 2003-10-15 | 中华电信股份有限公司 | Method and device for picking license plate area and correcting skew license plate from automobile image |
CN101021867A (en) * | 2007-03-22 | 2007-08-22 | 成都电子机械高等专科学校 | Image searching method based on image pyramid intermediate layer blocking |
CN101059870A (en) * | 2007-05-23 | 2007-10-24 | 华中科技大学 | Image cutting method based on attribute histogram |
-
2008
- 2008-07-14 CN CN 200810040548 patent/CN101630360B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1448885A (en) * | 2002-03-28 | 2003-10-15 | 中华电信股份有限公司 | Method and device for picking license plate area and correcting skew license plate from automobile image |
CN1424695A (en) * | 2002-12-24 | 2003-06-18 | 上海高德威智能交通系统有限公司 | Application method for badging image in expressway toll system |
CN101021867A (en) * | 2007-03-22 | 2007-08-22 | 成都电子机械高等专科学校 | Image searching method based on image pyramid intermediate layer blocking |
CN101059870A (en) * | 2007-05-23 | 2007-10-24 | 华中科技大学 | Image cutting method based on attribute histogram |
Non-Patent Citations (1)
Title |
---|
孙玉秋等.基于图像金字塔的分维融合算法.《计算机应用》.2005,第25卷(第5期),1064,1065,1075. * |
Also Published As
Publication number | Publication date |
---|---|
CN101630360A (en) | 2010-01-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN101630360B (en) | Method for identifying license plate in high-definition image | |
Badr et al. | Automatic number plate recognition system | |
CN102768731A (en) | Method and system for automatic positioning and identifying target based on high definition video images | |
US9818301B2 (en) | Lane correction system, lane correction apparatus and method of correcting lane | |
CN103116751B (en) | A kind of Method of Automatic Recognition for Character of Lcecse Plate | |
US7209832B2 (en) | Lane recognition image processing apparatus | |
CN101030256B (en) | Method and apparatus for cutting vehicle image | |
US8204305B2 (en) | Daytime and nighttime image recognizing method and apparatus | |
CN102799862A (en) | System and method for pedestrian rapid positioning and event detection based on high definition video monitor image | |
CN109255350B (en) | New energy license plate detection method based on video monitoring | |
CN105809184B (en) | Method for real-time vehicle identification and tracking and parking space occupation judgment suitable for gas station | |
KR101717613B1 (en) | The moving vehicle detection system using an object tracking algorithm based on edge information, and method thereof | |
CN107301405A (en) | Method for traffic sign detection under natural scene | |
KR100975749B1 (en) | Method for recognizing lane and lane departure with Single Lane Extraction | |
CN106682586A (en) | Method for real-time lane line detection based on vision under complex lighting conditions | |
CN109800752B (en) | Automobile license plate character segmentation and recognition algorithm based on machine vision | |
CN103824066A (en) | Video stream-based license plate recognition method | |
CN107563330B (en) | Horizontal inclined license plate correction method in surveillance video | |
CN105741559A (en) | Emergency vehicle lane illegal occupation detection method based on lane line model | |
CN102902951A (en) | System and method for vehicle target location and event detection on basis of high-definition video monitoring images | |
CN109241920B (en) | Lane line detection method for vehicle-mounted road monitoring and evidence obtaining system | |
CN105069456A (en) | License plate character segmentation method and apparatus | |
CN107563331B (en) | Road sign line detection method and system based on geometric relationship | |
CN112070039B (en) | Hash code-based vehicle collision detection method and system | |
CN111652033A (en) | Lane line detection method based on OpenCV |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20121219 Termination date: 20170714 |
|
CF01 | Termination of patent right due to non-payment of annual fee |