CN103310435A - Method for partitioning number plate characters by combining vertical projection and optimal path - Google Patents
Method for partitioning number plate characters by combining vertical projection and optimal path Download PDFInfo
- Publication number
- CN103310435A CN103310435A CN2012100759802A CN201210075980A CN103310435A CN 103310435 A CN103310435 A CN 103310435A CN 2012100759802 A CN2012100759802 A CN 2012100759802A CN 201210075980 A CN201210075980 A CN 201210075980A CN 103310435 A CN103310435 A CN 103310435A
- Authority
- CN
- China
- Prior art keywords
- character
- characters
- license plate
- vertical projection
- optimal path
- 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
Images
Landscapes
- Character Input (AREA)
Abstract
The invention provides a method for partitioning number plate characters by combining vertical projection and an optimal path. The method sequentially comprises the following steps: preprocessing a number plate image so as to obtain a character edge sharpen image of a number plate; performing rough partitioning on the number plate characters in the character edge sharpen image by adopting a vertical projection method so as to obtain partitioning point positions among character blocks; establishing a weight image on the basis of the character edge sharpen image; estimating the character width and character pitch of the number plate as well as the step size in search by utilizing the structure proportion character of the number plate and the rough partitioning result; calculating a partitioning path between two characters in an adhered character block in the weight image by adopting an A* search algorithm. According to the method provided by the invention, firstly, rough partitioning is performed on the number plate characters by adopting the vertical projection method, and then further partitioning is performed on the adhered characters by using the optimal path method, so that the character adhesion problem is solved, the character partitioning accuracy is improved, the method has the advantages of high flexibility and good adaptability, and is applicable to automatic number plate recognition systems in intelligent transportation systems of urban roads.
Description
Technical field
The present invention relates to a kind of registration number character dividing method, especially a kind of with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart.
Background technology
In intelligent transportation system, automatic license plate identification system is an important subsystem, occupies very important effect in the management of projects such as highway, charge station and parking lot.The main task of automatic license plate identification system is by analyzing, handle the vehicle image of gathering, by technological means such as computer vision, pattern-recognition and image processing vehicle license being identified automatically.It mainly by car plate location, characters on license plate is cut apart forms with the character recognition three parts, wherein to cut apart be condition precedent and the committed step of character recognition to characters on license plate, the accuracy rate of Character segmentation has material impact to the correct recognition rata of whole LPR system.
Common registration number character dividing method mainly contains vertical projection method, template matching method, based on the method for cluster analysis, these methods have relative merits separately.
1, based on the registration number character dividing method of vertical projection
Based on the registration number character dividing method of vertical projection, at first calculate the projection of licence plate character after the binaryzation, utilize the minimal value of projection histogram to find the segmentation candidates point, come separating character according to the prioris such as proportionate relationship of character fixed width, spacing again.
It is simple that this method has program design, the advantage that execution speed is fast, and for the measured License Plate Image of matter, it is very accurate to cut apart.But when noise seriously causes character adhesion or fracture, be easy to take place wrong branch problem.
2, based on the characters on license plate partitioning algorithm of template matches
Based on the registration number character dividing method of template matches, utilized characters on license plate height and character string structure proportion relation, at first according to the characters on license plate height, determine the size of other parts, set up character string and cut apart template, as shown in Figure 1.Then template is slided in image from left to right, add up character pixels point or number of edge points n in all character zones of each position simultaneously respectively
1With character pixels point or the number of edge points n in all zones, character pitch
2, and calculate its difference n
1-n
2At last, the n that compares each position
1-n
2, select n
1-n
2Template position when getting maximal value is the accurate position of characters on license plate string, thereby realizes cutting apart of characters on license plate.
The interference that can overcome character adhesion and rivet etc. behind the image binaryzation of this method, self-adaptation is strong.But its accuracy depends on the character duration of binaryzation effect and estimation strongly, and when distortion took place car plate, the charcter topology proportionate relationship can change, and the Character segmentation accuracy rate of this method will reduce greatly.
3, based on the registration number character dividing method of cluster analysis
Registration number character dividing method based on cluster analysis, each pixel in the license plate image behind the removal frame is carried out cluster by distance, at first remove the noise class that does not meet the character height feature, if remaining class then divides processing to the class of maximum less than 7, if remaining class is greater than 7, because the spacing between the character has equidistant character, get 6 class spacings successively, calculate variance, corresponding 7 classes of 6 distances of variance minimum are exactly character type.
This method can solve preferably that Chinese character is not communicated with, exist in the Character segmentation noise, car plate wearing and tearing cause difficult problems such as character adhesion; But the cluster process calculated amount is big, is not suitable for the application of real-time system, and to the easy cluster mistake of division character.
Except said method, registration number character dividing method also has the method analyzed based on connected domain and based on the method for mathematical morphology etc.Yet, be subjected to character picture that noise pollution obtains with traditional Character segmentation method when serious and often cut apart failure or mistake occurs and cut apart when license plate image exists, cause character recognition to be failed.
Summary of the invention
At the weak point of above-mentioned technology, the invention provides a kind of character adhesion problems that solves, improve the Character segmentation accuracy with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart.
For achieving the above object, the invention provides a kind of with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, may further comprise the steps:
(1) license plate image is carried out pre-service with the clear figure of the character edge that obtains car plate;
(2) adopt vertical projection method that the characters on license plate among the clear figure of character edge is carried out coarse segmentation, to obtain the cut-point position between the character block;
(3) set up weights figure on the basis of the clear figure of character edge;
(4) utilize car plate structure proportion feature and coarse segmentation result, estimate characters on license plate width and character pitch and step-size in search;
(5) adopt A
*Search algorithm calculates in the adhesion character block split path between two characters in weights figure.
In step (1), license plate image is carried out pre-service comprise license plate image is carried out gray processing processing, image enhancement processing and adopts sharp edge detection method computed image edge.
In step (2), after the projection threshold value is made as 0, in the clear figure of character edge, utilize the proportion structure relation of car plate height and characters on license plate string, the qualified character block of width also synthesizes a character block to form an adhesion character block with ineligible character block among the reservation coarse segmentation result, obtains the cut-point position between the character block.
In step (4), weights figure is formed after being endowed weights by each pixel among the clear figure of character edge, the cost that these weights need be paid during by this pixel as the path.
In step (5), calculate the cut-point between per two characters in the adhesion character block according to the step-size in search of estimating, the coboundary point with each cut-point column in weights figure is defined as starting point, and according to A
*Search algorithm calculates one and arrives the optimal path of lower boundary by starting point, with optimal path as the split path between two characters.
Compared with prior art, the present invention has the following advantages:
The registration number character dividing method that the present invention proposes at first adopts vertical projection method that the clear figure of the character edge of car plate is carried out coarse segmentation, re-using the optimal path method realizes further cutting apart to the adhesion character, with can better solve the character adhesion problems after existing Character segmentation method is compared, to improve the accuracy that characters on license plate is cut apart, has the dirigibility height, the advantage that adaptability is good, can satisfy the real-time requirement of automatic license plate identification system, be applicable to the application of automatic license plate identification system in the urban road intelligent transportation system.
The present invention adopts vertical projection method that characters on license plate is carried out coarse segmentation, can reduce the execution time of algorithm, and can provide reference by location accurately for the estimation that optimal path is cut apart starting point; In weight map, according to A
*The rule that search algorithm is followed is sought an optimal path and is realized cutting apart of character between per two characters, this optimal path is by the interval region of character, method by optimum route search realizes will obtaining a curve that just passes the zone, character pitch cutting apart of character, thereby guarantees correctness and the accuracy of Character segmentation.
Description of drawings
Fig. 1 is characters on license plate height and character string structure proportion graph of a relation;
The process flow diagram of Fig. 2 dividing method of the present invention;
Fig. 3 a is former license plate image;
The clear figure of character edge of the car plate that Fig. 3 b obtains after step of the present invention (1) is handled for Fig. 3 a;
Fig. 3 c is the image of another former car plate;
The clear figure of character edge of the car plate that Fig. 3 d obtains after step of the present invention (1) is handled for Fig. 3 c;
The exemplary plot that Fig. 4 a obtains after step of the present invention (2) is handled for Fig. 3 b;
The exemplary plot that Fig. 4 b obtains after step of the present invention (2) is handled for Fig. 3 d;
Fig. 5 is the exemplary plot of weights figure assignment rule;
The exemplary plot that Fig. 6 a and Fig. 6 b obtain after step of the present invention (5) is handled for Fig. 3 d;
Fig. 7 a to Fig. 7 d carries out the figure as a result of Character segmentation for adopting vertical projection method to serious car plate affected by noise;
Fig. 8 a to Fig. 8 d carries out the figure as a result of Character segmentation for adopting the inventive method to serious car plate affected by noise;
Fig. 9 a to Fig. 9 b carries out the figure as a result of Character segmentation for adopting vertical projection method to the serious car plate of quality degradation;
Figure 10 a to Figure 10 d carries out the figure as a result of Character segmentation for using the inventive method to the serious car plate of quality degradation.
Embodiment
Below in conjunction with drawings and Examples the present invention is described in further detail.
As shown in Figure 1, the invention provides a kind of with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, may further comprise the steps:
(1) license plate image is carried out pre-service with the clear figure of the character edge that obtains car plate;
Be specially, at first colored license plate image is carried out pre-service such as gray processing and figure image intensifying, then, in order to suppress noise, to obtain accurately continuous character edge, reduce the influence of the edge gray-scale map being carried out losing in the binaryzation process efficient frontier information simultaneously, use adaptive threshold, adopt edge clear detection method edge calculation figure, with the clear figure of the character edge that obtains car plate, it is good that the clear figure of this character edge has noise suppression effect, and rim detection is advantage accurately.The clear figure of this character edge searches for the preparation that optimal path carries out for the foundation of weights figure and in weights figure.Shown in Fig. 3 a to Fig. 3 d, be respectively the clear figure of character edge of the car plate that obtains after two former car plates are handled by step (1).
(2) adopt vertical projection method that the characters on license plate among the clear figure of character edge is carried out coarse segmentation, to obtain the cut-point position between the character block;
Be specially, in the clear figure of character edge, use vertical projection method that characters on license plate is carried out coarse segmentation, and the projection threshold value is made as 0.Utilize the proportion structure relation of car plate height and characters on license plate string, keep the qualified character block of width among the coarse segmentation result, and ineligible character block synthesized a character block, thus form an adhesion character block, obtain the cut-point position between the character block.
Fig. 4 a and Fig. 4 b are respectively the exemplary plot that Fig. 3 b and Fig. 3 d obtain respectively after step of the present invention (2) coarse segmentation is handled.Use vertical projection method that characters on license plate among the clear figure of character edge is carried out coarse segmentation, can reduce the execution time of algorithm, utilize the coarse segmentation result simultaneously as a reference, more accurately estimate the cut-point position between the adhesion character, with the accuracy of the characters on license plate segmentation result that improves optimum route search.
(3) set up weights figure on the basis of the clear figure of character edge;
Be specially, weights figure is formed after being endowed weights by each pixel among the clear figure of character edge, the cost that these weights need be paid during by this pixel as the path.
In the present embodiment, for the search optimal path to realize cutting apart of characters on license plate, based on the clear figure of character edge, and set up weight map according to following rule.In weight map, the cost that need pay when the value at each pixel place all represents certain paths by this, it is more big to be worth more big cost, and optimal path should be the path of total cost minimum.As shown in Figure 5, suppose constant A, B, C, D, E, its pass is A＞B＞C＞D 〉=E, then weights are given rule and are: marginal point is given maximum weights A, the territory of the neighbours up and down pixel of marginal point is given weights B, and neighbours territory, the left and right sides pixel of marginal point is given weights C, if desired, the neighbours territory pixel of the neighbours territory pixel of marginal point is given weights D, and other non-marginal points are given minimum weights E.In this example, A=20, B=5, C=3, D=E=1.
(4) utilize car plate structure proportion feature and coarse segmentation result to estimate character duration and character pitch, and calculate step-size in search according to character duration and character pitch;
Be specially, utilize car plate structure proportion feature and coarse segmentation result to estimate character duration W and character pitch D, estimate the cut-point column j between the adhesion character.Wherein,
Character duration is defined as W, and character pitch is defined as D, and calculates step-size in search step according to following formula:
Step=W+D/2, formula (1)
In addition, owing to when the search optimal path, centered by the cut-point column, may need to search for to the left and right sides, to walk around the edge pixel point, therefore, tolerance width definition dw is:
Dw=D/2, formula (2).
(5) adopt A
*Search algorithm obtains two split paths between the character in weights figure.
Be specially, the coboundary point of each cut-point column among the weights figure is defined as starting point, and according to A
*Search algorithm calculates one and arrives the optimal path of lower boundary by starting point, with optimal path as the split path between two characters.
Because the car plate left margin often is not equal to the left margin of first character, the right margin of car plate is not equal to the right margin of last character.For the accuracy that guarantees that cut-point is estimated, at first to guarantee the accurate of reference position.Therefore, according to order from left to right, be different for the processing of leftmost adhesion character block and the processing of other adhesion character block, concrete processing mode is described below:
(a) if working as the adhesion character block of pre-treatment is first character block, then the right margin right with this character block is reference, estimates first effective cut-point column x left,
Wherein x=right-step changes (3) over to.
(b) if when the adhesion character block of pre-treatment be not first character block, then with the left margin left of this character block as a reference, estimate first effective cut-point column x to the right,
Wherein x=left+step changes (4) over to.
(c) with (0, x) be starting point, calculate an optimal path that arrives lower boundary as the split path between two characters according to (5), note Far Left pixel column lx and rightmost pixel column rx on this path.Upgrade x according to following formula, repeat (3) up to x＜0:
X=(x+lx+rx)/3-step, formula (3)
(d) with (0, x) be starting point, calculate an optimal path that arrives lower boundary as the split path between two characters according to (5), note Far Left pixel column lx and rightmost pixel column rx on this path.Upgrade x according to following formula, repeat (4) up to x＞width-1, wherein width is the width of this character block:
X=(x+lx+rx)/3+step, formula (4)
According to A
*The thought of search algorithm, the present invention searches the method that optimal path abides by and is: starting point is defined as G, and G=0, the height of character block is defined as H, and H=Height, be i.e. the origin-to-destination vertical range of being expert at.Starting point moved to close in the tabulation, the neighbours territory pixel of starting point is added open tabulation, the father node that records these new interpolation nodes is starting point, and calculates their G and H value.The ranks coordinate of supposing newly to add certain pixel in the node for (i, j), then:
For each new interpolation node calculates F value (F=G+H), select the node of F minimum as present node, tabulation is closed in its adding, and with in its neighbours territory pixel or not in closing tabulation join and open tabulation, the father node that records these new interpolation pixels is present node, and calculates their G and H value according to formula (5).Repeat this process and equal Height up to the row i that certain is added into the pixel of closing tabulation.The exemplary plot that Fig. 6 a and Fig. 6 b obtain after step of the present invention (5) is handled for Fig. 3 d.
Serious car plate affected by noise is carried out the figure as a result of Character segmentation as Fig. 7 a to Fig. 7 d for adopting vertical projection method.Fig. 8 a to Fig. 8 d carries out the figure as a result of Character segmentation for adopting the inventive method to serious car plate affected by noise.Fig. 9 a to Fig. 9 b carries out the figure as a result of Character segmentation for adopting vertical projection method to the serious car plate of quality degradation.Figure 10 a to Figure 10 d carries out the figure as a result of Character segmentation for using the inventive method to the serious car plate of quality degradation.
Some affected by noise or second-rate license plate images are tested, draw Character segmentation method that the present invention proposes and compare to have than traditional Character segmentation method and littler cut apart mortality and mistake is cut apart situation, can realize more accurately cutting apart of characters on license plate.
The present invention has proposed a kind of registration number character dividing method based on optimal path from the essence of Character segmentation.Be example with horizontal character, utilize and have a certain size characteristics at interval between the connectedness of character and the character, the essence of Character segmentation is to seek the straight line section between per two characters, each bar straight-line segment is through the interval region between the character, a series of characters in the character string picture can be separated, in order to realize cutting apart of character.So it is the key that guarantees the Character segmentation accuracy that cut-point is estimated correctness.When the license plate image mass ratio is better, use traditional Character segmentation method can obtain the border, the left and right sides of each character accurately.Yet when being subjected to the influence of factors such as noise or car plate frame to cause having the character adhesion problems in the binary map when license plate image, traditional Character segmentation method often is difficult to estimate accurately the position of cut-point.Accurately cut apart the difficulty that the adhesion character faces in order to solve, the present invention uses the method based on optimum route search to realize cutting apart of character, obtains a curve that just passes the zone, character pitch, thereby guarantees the accuracy of Character segmentation.A
*Search algorithm can obtain a heuritic approach that connects the optimal path with minimum cost of start node and end node as the most frequently used pathfinding algorithm.By A
*The thought of search algorithm changes into the pathfinding problem with the Character segmentation problem among the present invention: at the computed segmentation point x of the place, coboundary of license plate image, be starting point with x, according to A
*The rule that search algorithm is followed is calculated an optimal path that arrives lower boundary as the split path between two characters, and this path may be vertical line, oblique line or a curve.The method dirigibility height that the present invention proposes, adaptability is good, can better solve the character adhesion problems than traditional Character segmentation method, has higher car plate and cuts apart accuracy rate.In addition, the search of considering optimal path is more consuming time, and vertical projection method is simply quick, and can provide reference by location accurately for the estimation that optimal path is cut apart starting point, the present invention uses vertical projection method to realize the coarse segmentation of characters on license plate earlier, re-uses the optimal path method adhesion character is realized further cutting apart.
The above only is preferred embodiment of the present invention, and is in order to limit the present invention, within the spirit and principles in the present invention not all, any modification of doing, is equal to replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (5)
- One kind with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, it is characterized in that, may further comprise the steps:(1) license plate image is carried out pre-service with the clear figure of the character edge that obtains car plate;(2) adopt vertical projection method that the characters on license plate among the clear figure of character edge is carried out coarse segmentation, to obtain the cut-point position between the character block;(3) set up weights figure on the basis of the clear figure of character edge;(4) utilize car plate structure proportion feature and coarse segmentation result, estimate characters on license plate width and character pitch and step-size in search;(5) adopt A *Search algorithm calculates in the adhesion character block split path between two characters in weights figure.
- According to claim 1 with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, it is characterized in that, in step (1), license plate image is carried out pre-service comprise license plate image is carried out gray processing processing, image enhancement processing and adopts sharp edge detection method computed image edge.
- According to claim 1 with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, it is characterized in that, in step (2), after the projection threshold value is made as 0, in the clear figure of character edge, utilize the proportion structure relation of car plate height and characters on license plate string, keep the qualified character block of width among the coarse segmentation result, and ineligible character block synthesized a character block to form an adhesion character block, obtain the cut-point position between the character block.
- According to claim 1 with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, it is characterized in that, in step (4), weights figure is formed after being endowed weights by each pixel among the clear figure of character edge, the cost that these weights need be paid during by this pixel as the path.
- According to claim 1 with vertical projection and optimal path in conjunction with the method that characters on license plate is cut apart, it is characterized in that, in step (5), calculate the cut-point between per two characters in the adhesion character block according to the step-size in search of estimating, coboundary point with each cut-point column in weights figure is defined as starting point, and according to A *Search algorithm calculates one and arrives the optimal path of lower boundary by starting point, with optimal path as the split path between two characters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210075980.2A CN103310435B (en) | 2012-03-21 | 2012-03-21 | Combine vertical projection and optimal path the method split characters on license plate |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210075980.2A CN103310435B (en) | 2012-03-21 | 2012-03-21 | Combine vertical projection and optimal path the method split characters on license plate |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103310435A true CN103310435A (en) | 2013-09-18 |
| CN103310435B CN103310435B (en) | 2016-04-06 |
Family
ID=49135612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210075980.2A Expired - Fee Related CN103310435B (en) | 2012-03-21 | 2012-03-21 | Combine vertical projection and optimal path the method split characters on license plate |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103310435B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104715252A (en) * | 2015-03-12 | 2015-06-17 | 电子科技大学 | License plate character segmentation method with combination of dynamic template and pixel points |
| WO2017071063A1 (en) * | 2015-10-30 | 2017-05-04 | 小米科技有限责任公司 | Area identification method and device |
| CN106650732A (en) * | 2016-12-27 | 2017-05-10 | 深圳市捷顺科技实业股份有限公司 | License plate identification method and apparatus |
| CN107993130A (en) * | 2017-12-26 | 2018-05-04 | 北京三快在线科技有限公司 | Method for processing business, system and electronic equipment |
| CN108830278A (en) * | 2018-05-17 | 2018-11-16 | 河南思维轨道交通技术研究院有限公司 | A kind of character string picture recognition methods |
| CN108960239A (en) * | 2018-07-10 | 2018-12-07 | 武汉科技大学 | A kind of laser-induced thermal etching detonator with metal shell code character dividing method based on image procossing |
| CN109325492A (en) * | 2018-08-17 | 2019-02-12 | 平安科技（深圳）有限公司 | Character segmentation method, apparatus, computer equipment and storage medium |
| CN110443808A (en) * | 2019-07-04 | 2019-11-12 | 杭州深睿博联科技有限公司 | Medical image processing method and device, equipment, storage medium for the detection of brain middle line |
| CN111079762A (en) * | 2019-11-26 | 2020-04-28 | 合肥联宝信息技术有限公司 | Cutting method of adhesive characters and electronic equipment |
| CN111191550A (en) * | 2019-12-23 | 2020-05-22 | 初建刚 | Visual perception device and method based on automatic dynamic adjustment of image sharpness |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100218394B1 (en) * | 1997-03-12 | 1999-09-01 | 이종수 | Method for recognizing a license number for a car |
| CN101408933A (en) * | 2008-05-21 | 2009-04-15 | 浙江师范大学 | Method for recognizing license plate character based on wide gridding characteristic extraction and BP neural network |
| CN101510265A (en) * | 2009-03-23 | 2009-08-19 | 浙江大学 | Association moving number character intelligent identification method |
| CN102129560A (en) * | 2010-01-18 | 2011-07-20 | 富士通株式会社 | Method and device for identifying characters |
| CN102222226A (en) * | 2011-06-21 | 2011-10-19 | 杭州电子科技大学 | Priori analysis based iterative method for segmenting characters of licence plate |
-
2012
- 2012-03-21 CN CN201210075980.2A patent/CN103310435B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100218394B1 (en) * | 1997-03-12 | 1999-09-01 | 이종수 | Method for recognizing a license number for a car |
| CN101408933A (en) * | 2008-05-21 | 2009-04-15 | 浙江师范大学 | Method for recognizing license plate character based on wide gridding characteristic extraction and BP neural network |
| CN101510265A (en) * | 2009-03-23 | 2009-08-19 | 浙江大学 | Association moving number character intelligent identification method |
| CN102129560A (en) * | 2010-01-18 | 2011-07-20 | 富士通株式会社 | Method and device for identifying characters |
| CN102222226A (en) * | 2011-06-21 | 2011-10-19 | 杭州电子科技大学 | Priori analysis based iterative method for segmenting characters of licence plate |
Non-Patent Citations (2)
| Title |
|---|
| 马瑞，杨静宇: "一种利用灰度信息和二值信息的手写字符分割", 《计算机工程》, 31 August 2007 (2007-08-31) * |
| 马瑞: "非限制手写字符分割中相关技术与算法的研究", 《中国博士学位论文全文数据库信息科技辑》, 15 December 2007 (2007-12-15) * |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104715252B (en) * | 2015-03-12 | 2018-05-18 | 电子科技大学 | A kind of registration number character dividing method of dynamic template combination pixel |
| CN104715252A (en) * | 2015-03-12 | 2015-06-17 | 电子科技大学 | License plate character segmentation method with combination of dynamic template and pixel points |
| WO2017071063A1 (en) * | 2015-10-30 | 2017-05-04 | 小米科技有限责任公司 | Area identification method and device |
| US10157326B2 (en) | 2015-10-30 | 2018-12-18 | Xiaomi Inc. | Method and device for character area identification |
| CN106650732B (en) * | 2016-12-27 | 2020-04-28 | 深圳市捷顺科技实业股份有限公司 | License plate recognition method and device |
| CN106650732A (en) * | 2016-12-27 | 2017-05-10 | 深圳市捷顺科技实业股份有限公司 | License plate identification method and apparatus |
| CN107993130A (en) * | 2017-12-26 | 2018-05-04 | 北京三快在线科技有限公司 | Method for processing business, system and electronic equipment |
| CN108830278A (en) * | 2018-05-17 | 2018-11-16 | 河南思维轨道交通技术研究院有限公司 | A kind of character string picture recognition methods |
| CN108830278B (en) * | 2018-05-17 | 2021-11-02 | 河南思维轨道交通技术研究院有限公司 | Character string image recognition method |
| CN108960239A (en) * | 2018-07-10 | 2018-12-07 | 武汉科技大学 | A kind of laser-induced thermal etching detonator with metal shell code character dividing method based on image procossing |
| CN108960239B (en) * | 2018-07-10 | 2021-02-19 | 武汉科技大学 | Image processing-based laser etching metal detonator coded character segmentation method |
| CN109325492A (en) * | 2018-08-17 | 2019-02-12 | 平安科技（深圳）有限公司 | Character segmentation method, apparatus, computer equipment and storage medium |
| CN109325492B (en) * | 2018-08-17 | 2023-12-19 | 平安科技（深圳）有限公司 | Character cutting method, device, computer equipment and storage medium |
| CN110443808A (en) * | 2019-07-04 | 2019-11-12 | 杭州深睿博联科技有限公司 | Medical image processing method and device, equipment, storage medium for the detection of brain middle line |
| CN110443808B (en) * | 2019-07-04 | 2022-04-01 | 杭州深睿博联科技有限公司 | Medical image processing method and device for brain midline detection, equipment and storage medium |
| CN111079762A (en) * | 2019-11-26 | 2020-04-28 | 合肥联宝信息技术有限公司 | Cutting method of adhesive characters and electronic equipment |
| CN111079762B (en) * | 2019-11-26 | 2022-02-08 | 合肥联宝信息技术有限公司 | Cutting method of adhesive characters and electronic equipment |
| CN111191550A (en) * | 2019-12-23 | 2020-05-22 | 初建刚 | Visual perception device and method based on automatic dynamic adjustment of image sharpness |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103310435B (en) | 2016-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103310435A (en) | Method for partitioning number plate characters by combining vertical projection and optimal path | |
| CN104008645B (en) | One is applicable to the prediction of urban road lane line and method for early warning | |
| CN109459045B (en) | Improved interactive voting matching method for low-frequency GPS track | |
| CN102509291B (en) | Pavement disease detecting and recognizing method based on wireless online video sensor | |
| CN104463903A (en) | Pedestrian image real-time detection method based on target behavior analysis | |
| CN104318258A (en) | Time domain fuzzy and kalman filter-based lane detection method | |
| JP2016075905A5 (en) | ||
| CN104200485A (en) | Video-monitoring-oriented human body tracking method | |
| CN108961758B (en) | Road junction widening lane detection method based on gradient lifting decision tree | |
| CN104599502A (en) | Method for traffic flow statistics based on video monitoring | |
| CN111341103B (en) | Lane information extraction method, device, equipment and storage medium | |
| CN108828608B (en) | Laser radar background data filtering method in vehicle detection method | |
| CN103268489A (en) | Motor vehicle plate identification method based on sliding window searching | |
| CN101329815A (en) | Novel system and method for detecting four-phase vehicle flow of a traffic road junction | |
| CN105096315A (en) | Method for segmenting heterogeneous super-pixel SAR (Synthetic Aperture Radar) image based on Gamma distribution | |
| CN107909047A (en) | A kind of automobile and its lane detection method and system of application | |
| CN103077531A (en) | Grayscale target automatic tracking method based on marginal information | |
| CN108805312B (en) | Method and device for determining adjacent bayonets | |
| CN106845482A (en) | A kind of license plate locating method | |
| CN106227859A (en) | The method identifying the vehicles from gps data | |
| CN113127466B (en) | Vehicle track data preprocessing method and computer storage medium | |
| CN102867415A (en) | Video detection technology-based road jam judgement method | |
| CN103794050A (en) | Real-time transport vehicle detecting and tracking method | |
| CN102156989B (en) | Vehicle blocking detection and segmentation method in video frame | |
| CN104599291A (en) | Structural similarity and significance analysis based infrared motion target detection method |
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: 20160406 Termination date: 20190321 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |