CN102722707A - License plate character segmentation method based on connected region and gap model - Google Patents
License plate character segmentation method based on connected region and gap model Download PDFInfo
- Publication number
- CN102722707A CN102722707A CN2012101898982A CN201210189898A CN102722707A CN 102722707 A CN102722707 A CN 102722707A CN 2012101898982 A CN2012101898982 A CN 2012101898982A CN 201210189898 A CN201210189898 A CN 201210189898A CN 102722707 A CN102722707 A CN 102722707A
- Authority
- CN
- China
- Prior art keywords
- character
- license plate
- connected region
- gap
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Character Input (AREA)
Abstract
The invention belongs to the technical field of digital image processing and pattern recognition, and particularly relates to a license plate character segmentation method based on a connected region and a gap model. Firstly, a license plate gray image is subjected to binarization, then a connected region is marked in the binarized license plate image, next a character region in a license plate is acquired initially, finally a gap model is established to confirm the position of each character of the license plate according to proportions among characters of a standard license plate, and accordingly segmentation of the character region in the license plate is achieved. The characters of the standard license plate and size of gaps are utilized effectively, and influence of illegible characters caused by stain of the license plate, lighting of the image and the like on the character segmentation is avoided. Compared with a traditional license plate character segmentation method based on a connected region and projection and segmentation, the license plate character segmentation method based on the connected region and the gap model can improve accuracy of license plate character segmentation effectively and well solve the problem of license plate character segmentation caused by the illegible image.
Description
Technical field
The invention belongs to Digital Image Processing and mode identification technology, be specifically related to the method for traffic scene car plate identification in the intelligent traffic administration system.
Technical background
Intelligent transportation system (Intelligent Transportation System, be called for short ITS) is the system ensemble with various traffic route monitor management system, vehicle control system and the highway traffic safety system of the integrated actual demand that applies to ground transport of electronic technology, computer technology, infotech, sensor technology and system engineering technology.Car plate identification (Vehicle License Plate Recognition is called for short VLPR) is one of important component part of intelligent transportation system, and its application is very extensive.Car plate identification is the basis with technology such as Digital Image Processing, pattern-recognition, computer visions, and the vehicle image or the video sequence of shot by camera are analyzed, and obtains the unique number-plate number of each automobile, thereby accomplishes identifying.Through some follow-up processing means; License plate recognition technology can be realized functions such as parking lot fee collection management, the measurement of magnitude of traffic flow controlling index, vehicle location, automobile burglar, highway hypervelocity robotization supervision; To safeguarding traffic safety and urban public security; Prevent traffic jam, realize that the automatic management of traffic has realistic meanings.
License plate recognition technology mainly comprises three steps: the location of license plate image, characters on license plate cut apart the identification with characters on license plate." car plate location " is to utilize image processing method and mode identification technology to have the position of accurately orienting car plate the vehicle number word image of uncertain background from a width of cloth, and cutting apart and discern to handle for follow-up characters on license plate provides data source accurately and reliably." Character segmentation " is that license plate area is divided into single character zone, is used for next step character recognition.The character of car plate is made up of limited Chinese character, English alphabet and numeral, " identification of characters on license plate " thus then be the character to be identified that splits to be classified discern them.
Wherein, " Character segmentation " is the ring that license plate recognition technology is taken over from the past and set a new course for the future, and the correctness of Character segmentation directly has influence on the result of Recognition of License Plate Characters.But " Character segmentation " but is the bigger part of difficulty, and its difficulty is mainly reflected on the following problem:
(1) because weather condition; The license plate image clean-up performance differs, and in the car plate bianry image that obtains after the car plate pre-service, possibly have the character adhesion; Also possibly there is the character stroke fracture; Therefore, in the Character segmentation method, should avoid a plurality of Character segmentation is a character, and a Character segmentation is become the maloperation of a plurality of characters.
(2) because the width of character " 1 " is narrow than the width of other character, how to distinguish character " 1 ' ' with the interference stroke, the vertical frame that particularly remains about car plate should take into full account in method.
(3) through pretreated car plate binary map; Can not accurately navigate to the up-and-down boundary of character, also receive the interference of frame and rivet, also can there be interference in the right and left of car plate in addition; The frame of the right and left particularly, and possibly stick together with first or last character.Therefore the dividing method that is designed can not receive the influence of these interference regions.
The influence of illumination during (4) owing to shooting; The all or part of zone of car plate seems bright especially; Cause character and background color discrimination little, dividing method should consider that a kind of suitable gray processing and binarization method distinguish background area and character zone accurately.
Several kinds of Character segmentation methods commonly used at present have:
(1) based on the Character segmentation method of sciagraphy.Utilize vertical projection method to find the best cutting point between the character fast, and transverse projection capable of using remove interference such as car plate frame.
(2) based on the separating character method of characters on license plate geometric properties.This method is carried out a series of morphological operations with the license plate image of mathematical morphology after to binaryzation earlier, gets rid of some garbages, makes that the space becomes greatly between character and car plate left and right side frame, character and the character, is convenient to mark intercharacter vertical divider.
(3) based on the maximum between-cluster variance registration number character dividing method of template matches.According to the structure and the size characteristic of character string, designed characters on license plate string template, this template is slided to mate at license plate area and is classified, and combines maximum between-cluster variance decision rule to confirm best match position, cuts apart characters on license plate.
(4) based on the Character segmentation method of scanning, this method utilization is confirmed the up-and-down boundary of character by the centre to the method for two ends search, and utilizes the dependency rule of one dimension circulation zero clearing method and characters on license plate to carry out vertical segmentation to obtain single character.
(5) cut apart the method for characters on license plate based on connected component analysis.Promptly constitute the principle of a connected region, combine the dependency rule of characters on license plate again, thereby better solved the character cutting problem of license plate under complex background condition according to the pixel that belongs to same character.
(6) based on the registration number character dividing method of neural network and color characteristic.In color space, utilize tree type judgement structure, at first identify white, black picture element in the license plate area, utilize network that the indigo plant in the license plate area, red, yellow and other colors are discerned then by monochrome information.According to the color characteristic of car plate, after judging the type of car plate, license plate area is carried out binary conversion treatment.After removing car plate frame and rivet, the connectedness of comprehensive utilization sciagraphy and character is cut apart characters on license plate.
More than these registration number character dividing methods; It is the method that improves gradually in the registration number character dividing method evolution; But all do not have well to solve stained or receive to have in the Character segmentation, particularly car plate of illumination effect car plate the situation that two or three character zones are fuzzy or obscure, these methods often can accurately not cut apart; In the actual traffic scene, the accuracy rate of Character segmentation will reduce significantly.
Because the complicacy of field of traffic application scenarios, license plate recognition technology problem to be solved are quite complicated.The present license plate recognition technology part that in practical application, also comes with some shortcomings, for example: extraneous brightness is low excessively, illumination condition abominable, special weather condition, complicated non-license plate area interference etc. cause certain degree of difficulty all can for the location of car plate; The meeting that influences of the cleanliness of car plate self or illumination condition causes difficulty to character cutting; The identification of similar character is distinguished and is also had certain difficulty.
Summary of the invention
The objective of the invention is, above-mentioned car plate is stained spares the problem that causes characters on license plate to be difficult to distinguish with uneven illumination in order to solve, and a kind of registration number character dividing method based on connected region and gap model is provided.
The registration number character dividing method based on connected region and gap model that the present invention proposes is after the location obtains car plate, carries out the operation of following steps:
(1) gray processing and binary conversion treatment
1. adopt the component method that license plate image is carried out gray processing and handle, the steps include: the gray-scale value of the three-component brightness of the RGB in the coloured image as three gray level images chosen wherein a kind of gray level image according to application need;
Because the red component of blue car plate is apparent in view, the green component of yellow car plate is apparent in view, in order to increase the discrimination between character zone and the background, need carry out gray processing to blue and yellow car plate and handle;
2. after the gray processing of step (1)-is 1. handled; License plate image is carried out binary conversion treatment, the steps include: to adopt NiBlack dynamic threshold binaryzation algorithm, calculate certain pixel (coordinate points) threshold value in the license plate image through formula; The pixel value of this threshold value and coordinate points is done comparison; Less than threshold value is background color, is character color greater than threshold value, obtains the car plate bianry image;
(2) mark connected region
Said connected region is eight connected regions, promptly arbitrary pixel with it around the pixel value of adjacent eight pixels when identical, then be regarded as being communicated with between them; As standard, with the bianry image of step (1)-2. obtain method mark connected region according to recurrence; After connected region is confirmed through mark, minimum with the maximum horizontal ordinate of record connected region;
(3) remove the impurity noise
Through the geometric properties of minimum with the maximum horizontal ordinate calculating connected region of the connected region of step (2) record, judge whether connected region is character zone; Then keep if satisfy condition; Otherwise, then be made as background colour;
(4) transverse cuts
The car plate of vehicle is fixed on the vehicle through rivet, and the part character zone links to each other with rivet in the bianry image of car plate, and it is communicated with frame becomes a connected region, and such zone will be deleted because of the geometric properties that does not meet character zone; Upper and lower two centre coordinate points of the character zone that remains exist respectively in the set, and these two coordinate point sets are carried out linear fit respectively, the upper and lower straight line that respectively obtains; Article two, straight line is separated character zone and rivet, repeats the operation of step (2) and step (3), and the character zone that will link to each other with rivet remains;
(5) gap is calculated and is recovered the residue character.
Through calculating the gap between adjacent connected region, calculate between these two connected regions and have several characters and several gap, again according to the normal width of known region with calculate the gap width that obtains and recover the character zone deleted by mistake; Normal width and calibrated altitude through these continuous reserved character zone calculating characters; Through the width of normal width calculated gap, utilize the character frame of confirming by normal width and calibrated altitude at last again, expand to the left and right by the gap; Final seven character zones confirming car plate, completion is cut apart.
Good effect of the present invention is:
(1) utilizes the characteristic of bianry image character connected sum characters on license plate standard geometrical shapes, can remove irregular impurity noise through the geometry information of connected region.
(2) utilize the positional information of fixed character zone to carry out upper and lower cutting, can well avoid impurity such as rivet to link to each other to the influence of cutting apart with character.
(3) utilize the ratio of car plate standard character size and inter-character space, the position at calculating character place, the effect of key does not meet the character zone that geometrical property is deleted with regard to being to recover owing to spot or illumination effect.
Description of drawings
Fig. 1 is the FB(flow block) that the present invention is based on the registration number character dividing method of connected region and gap model.
Fig. 2 is NiBlack binaryzation algorithm principle figure.
Fig. 3 is a mark circulation region method process flow diagram.
Fig. 4 removes the impurity method flow diagram.
Fig. 5 is the type list of characters on license plate zone reserve area.
Fig. 6 is standard automobile car plate (image derives from the GA36-2007 standard).
Fig. 7 is a characters on license plate gap-type table.
Fig. 8 is the type list after the expansion of car plate bianry image reserve area gap.
Fig. 9 is each step design sketch that the inventive method is handled a car plate.
Embodiment
The present invention is based on the embodiment of the registration number character dividing method of connected region and gap model below in conjunction with explained, but be noted that enforcement of the present invention is not limited to following embodiment.
A kind of registration number character dividing method based on connected region and gap model; At first license plate image is carried out gray processing and binaryzation; Remove the impurity noise after the mark connected region; Then after the transverse cuts again the repeating label connected region with remove two steps of impurity noise, calculate through the gap at last and recover the residue character and obtain seven character zones.
The concrete operation step of the inventive method is shown in accompanying drawing 1.
One, gray processing and binary conversion treatment
At first, this is blue car plate, carries out the red component binaryzation; Carry out NiBlack dynamic threshold binaryzation then.
Confirm the threshold value of this point through the statistical property (average and variance) of each its neighborhood of pixel in the computed image, the method schematic diagram is as shown in Figure 2, shown in the following formula of computing method:
Formula (1) is threshold value
computing formula of coordinate points in the image
; K is a weight coefficient,
,
be respectively to be the variance and the average of all pixels in the foursquare neighborhood in center with coordinate points
.After obtaining the threshold value of this point, just the pixel value of itself and this coordinate points being done comparison, is background color less than threshold value T, is character color greater than threshold value T.Shown in the following formula:
The pixel value of
coordinate points in
the expression bianry image
in the formula (2);
is the pixel value of coordinate points in the gray level image
, the threshold value that
calculates for above-mentioned steps.
Can find out that because the pollution of car plate is more, the license plate image extrinsic region after the binaryzation is a lot, noise effect is very big.
Two, mark connected region
According to the result of binaryzation in a last step, black is character color, and white is background color.According to the thought mark connected region of recurrence, process flow diagram is as shown in Figure 3:
Step 1. is provided with mark to each pixel, initially all is made as 0;
Each pixel of step 2. traversal is if pixel value is black and is labeled as 0 that counter adds 1, if having traveled through all pixels then finishes, otherwise carry out step 3;
This point of step 3. mark is the value of current counter, and upgrades the minimum of this connected region and the value of maximum horizontal ordinate;
The eigenwerts such as height, width and depth-width ratio that step 5. is calculated this connected region according to the minimum and the maximum horizontal ordinate of record, and judge according to these eigenwerts whether this connected region is character zone, if then keep this zone; If not, be background colour then with this zone marker, counter subtracts 1.Return and continue execution in step 2.
After connected region is confirmed through mark, the minimum and the maximum horizontal ordinate of record connected region.
Three, press geometric properties and remove the impurity noise region
Owing to the reasons such as spot on illumination, car plate frame and the car plate; License plate image impurity noise after the binaryzation is more, judges through the minimum of connected region and the geometric properties of maximum horizontal ordinate calculating impurity noise whether the connected region that the back mark comes out is character zone.If satisfy condition, then keep; Otherwise, then be made as background colour.It is as shown in Figure 4 to remove the impurity program circuit; Calculate the height of connected region respectively, the ratio of width to height of the width of connected region, the angle of connected region, connected region judges whether these geometric properties meet the condition of character zone; Keep qualified connected region,, then delete this connected region if do not satisfy condition; The connected region that calculate to keep is again calculated the mean breadth and the average height of connected region, and whether the geometric properties of judging connected region is again compared with mean value to have more greatly and departed from, if then delete this connected region.After this step operation, only can see that the zone of " E " and " 4 " remains.
Four, cutting up and down
Minimum and maximum horizontal ordinate through the connected region that remains; The height and the positional information in calculating character zone; The last centre coordinate point of the character zone that remains is existed in the set; This coordinate point set is carried out linear fit through least square method, obtain the straight line that a match is come out; With identical method, the following central point of the character zone that remains is carried out linear fit; Article two, straight line cuts character up and down, and " B " and " 2 " and up and down rivet or impurity are separated; Again carry out the mark connected region then and remove these two steps of impurity, feasible " B " and " 2 " these two character zones also meet the geometric properties of character and remain.
Five, calculate in the gap
Through the calculating in several steps of front, because the influence of spot and the impurity interference of noise such as fracture of Chinese character, not every character zone can both keep.In order to recover to remain character conveniently, wouldn't utilize the Chinese character zone, if can recover to remain 6 letter and numbers smoothly, the position of Chinese character can be confirmed through these letter and numbers.
The type list of characters on license plate reserve area is as shown in Figure 5; The character zone that solid black squares or circular expression keep; And the character zone that hollow square or circular expression are deleted by mistake, recovering hollow character zone is the precondition of correct separating character.
As shown in Figure 6; Since the standard car plate all follow unified standard: " 7 character strings forming by a province Chinese character heel character or arabic numeral; the array format of standard car plate is:
; wherein
is the abbreviation of each province, municipality directly under the Central Government;
is capitalization English letter; be blank character, and 5 characters of back are English character or arabic numeral.The shared width of each character is certain in the original size of car plate, is 45mm, and between the character and the gap between character and the blank character is 12mm, the wide 10mm of blank character, character height is 90mm ".So the ratio that the wide existence of standard clearance and standard word is fixing, the gap of wherein back five characters and preceding two characters and the wide ratio of standard word are about 0.267, are referred to as little gap; Leave out owing to blank character can be considered to impurity in said method,, be referred to as big gap so second and the 3rd intercharacter gap and the wide ratio of standard word are about 0.756.
Gap-type table as shown in Figure 7; Can be through calculating the gap between adjacent connected region; Calculate and have what characters and several big or small gap between these two connected regions; Normal width according to known region recovers to be made these connected region inside not have the character zone of disappearance by the character zone of deleting by mistake with the big closely spaced width that calculates again.Two closely spaced width of a character are satisfied in zone between " E " and " 4 ", so the character " 1 " in the middle of recovering.
After the expansion of the gap shown in the gap-type table; The character zone " B E142 " that keeps is continuous; The character zone disappearance on only remaining both sides is waited to recover, thus the type of reserve area will reduce a lot, as shown in Figure 8; When wherein remaining three continuous character zones, can whether be that unique remaining character zone in the left and right sides of confirming how to expand is come in big gap through judging first gap; Because " B " and " E " meets the width in a big gap, thus about respectively expand a character.In addition when 4 and 3 and even 2 the continuous connected regions that do not contain big gap exist, can judge that continuous connected region right side can expand several character zones to the width on border, expansion about carrying out with this.
Utilize these continuous reserved character zones then, the normal width of calculating character and calibrated altitude are confirmed the character frame, finally confine seven character zones of car plate, and completion is cut apart; The design sketch of a car plate of said method processing is as shown in Figure 9.
Claims (1)
1. the registration number character dividing method based on connected region and gap model is characterized in that, after the location obtains car plate, carries out the operation of following steps:
(1) gray processing and binary conversion treatment
1. adopt the component method that license plate image is carried out gray processing and handle, the steps include: the gray-scale value of the three-component brightness in the coloured image as three gray level images chosen wherein a kind of gray level image according to application need;
Blue and yellow car plate are carried out gray processing to be handled;
2. after the gray processing of step (1)-is 1. handled; License plate image is carried out binary conversion treatment, the steps include: to adopt NiBlack dynamic threshold binaryzation algorithm, calculate certain pixel threshold value in the license plate image; The pixel value of this threshold value and coordinate points is done comparison; Less than threshold value is background color, is character color greater than threshold value, obtains the car plate bianry image;
(2) mark connected region
Said connected region is eight connected regions, promptly arbitrary pixel with it around the pixel value of adjacent eight pixels when identical, then be regarded as being communicated with between them; As standard, with the bianry image of step (1)-2. obtain method mark connected region according to recurrence; After connected region is confirmed through mark, minimum with the maximum horizontal ordinate of record connected region;
(3) remove the impurity noise
Through the geometric properties of minimum with the maximum horizontal ordinate calculating connected region of the connected region of step (2) record, judge whether connected region is character zone; Then keep if satisfy condition; Otherwise, then be made as background colour;
(4) transverse cuts
The car plate of vehicle is fixed on the vehicle through rivet, and the part character zone links to each other with rivet in the bianry image of car plate, and it is communicated with frame becomes a connected region, and such zone will be deleted because of the geometric properties that does not meet character zone; Upper and lower two centre coordinate points of the character zone that remains exist respectively in the set, and these two coordinate point sets are carried out linear fit respectively, the upper and lower straight line that respectively obtains; Article two, straight line is separated character zone and rivet, repeats the operation of step (2) and step (3), and the character zone that will link to each other with rivet remains;
(5) gap is calculated and is recovered the residue character;
Through calculating the gap between adjacent connected region, judge between these two connected regions to have several characters and several gap that the normal width according to known region recovers by the character zone of deleting by mistake with the gap width of calculating acquisition again; Normal width and calibrated altitude through these continuous reserved character zone calculating characters; Through the width of normal width calculated gap, utilize the character frame of confirming by normal width and calibrated altitude at last again, expand to the left and right by the gap; Final seven character zones confirming car plate, completion is cut apart.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101898982A CN102722707A (en) | 2012-06-11 | 2012-06-11 | License plate character segmentation method based on connected region and gap model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2012101898982A CN102722707A (en) | 2012-06-11 | 2012-06-11 | License plate character segmentation method based on connected region and gap model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102722707A true CN102722707A (en) | 2012-10-10 |
Family
ID=46948457
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2012101898982A Pending CN102722707A (en) | 2012-06-11 | 2012-06-11 | License plate character segmentation method based on connected region and gap model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102722707A (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103390160A (en) * | 2013-07-24 | 2013-11-13 | 佳都新太科技股份有限公司 | License plate character segmentation algorithm based on connected domain and template |
| CN104200210A (en) * | 2014-08-12 | 2014-12-10 | 合肥工业大学 | License plate character segmentation method based on parts |
| CN104636748A (en) * | 2013-11-14 | 2015-05-20 | 张伟伟 | License plate recognition method and device |
| 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 |
| CN105678207A (en) * | 2014-11-19 | 2016-06-15 | 富士通株式会社 | Device and method for identifying content of target nameplate image from given image |
| CN106446898A (en) * | 2016-09-14 | 2017-02-22 | 宇龙计算机通信科技(深圳)有限公司 | Extraction method and extraction device of character information in image |
| CN106650741A (en) * | 2016-12-30 | 2017-05-10 | 深圳市捷顺科技实业股份有限公司 | License plate character recognition method and device |
| CN106709394A (en) * | 2016-12-12 | 2017-05-24 | 北京慧眼智行科技有限公司 | Image processing method and device |
| CN107273892A (en) * | 2017-06-12 | 2017-10-20 | 北京智芯原动科技有限公司 | A kind of registration number character dividing method and device |
| CN107392111A (en) * | 2017-06-27 | 2017-11-24 | 青岛海信电器股份有限公司 | Advertisement recognition method and device |
| CN107527056A (en) * | 2017-09-01 | 2017-12-29 | 南京邮电大学 | A kind of character segmentation method based on coarse positioning car plate |
| CN107610132A (en) * | 2017-08-28 | 2018-01-19 | 西北民族大学 | A kind of ancient books file and picture greasiness removal method |
| CN108256526A (en) * | 2017-12-07 | 2018-07-06 | 上海理工大学 | A kind of automobile license plate position finding and detection method based on machine vision |
| CN108734167A (en) * | 2018-05-08 | 2018-11-02 | 湖南开启时代电子信息技术有限公司 | A kind of contaminated film character recognition method |
| CN109034157A (en) * | 2017-06-08 | 2018-12-18 | 北京君正集成电路股份有限公司 | Licence plate recognition method and device |
| CN109086767A (en) * | 2018-07-11 | 2018-12-25 | 于洋 | The dividing method of fuzzy license plate in a kind of monitoring scene |
| CN109241975A (en) * | 2018-08-27 | 2019-01-18 | 电子科技大学 | A kind of registration number character dividing method based on character center point location |
| CN109271988A (en) * | 2018-08-30 | 2019-01-25 | 中国传媒大学 | A kind of subtitle extraction method based on image segmentation and dynamic threshold |
| CN109961553A (en) * | 2017-12-26 | 2019-07-02 | 航天信息股份有限公司 | Invoice number recognition methods, device and tax administration self-service terminal system |
| CN110728275A (en) * | 2018-07-16 | 2020-01-24 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and storage medium |
| CN111563402A (en) * | 2019-02-14 | 2020-08-21 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device, terminal and storage medium |
| CN112381088A (en) * | 2020-11-11 | 2021-02-19 | 上善智城(苏州)信息科技有限公司 | License plate recognition method and system for oil tank truck |
| CN112836696A (en) * | 2019-11-22 | 2021-05-25 | 搜狗（杭州）智能科技有限公司 | Text data detection method and device and electronic equipment |
| CN113096095A (en) * | 2021-04-12 | 2021-07-09 | 江苏理工学院 | Machine vision-based adhesive tablet counting method and counting system |
| CN114549811A (en) * | 2021-12-31 | 2022-05-27 | 哈尔滨理工大学 | License plate identification method based on license plate spacer body characteristics |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1085456A2 (en) * | 1999-09-15 | 2001-03-21 | Siemens Corporate Research, Inc. | Character segmentation method for vehicle license plate recognition |
| CN101154271A (en) * | 2006-09-30 | 2008-04-02 | 电子科技大学中山学院 | License plate character segmentation method based on fast area labeling algorithm and license plate large-spacing locating method |
| CN102184399A (en) * | 2011-03-31 | 2011-09-14 | 上海名图信息技术有限公司 | Character segmenting method based on horizontal projection and connected domain analysis |
-
2012
- 2012-06-11 CN CN2012101898982A patent/CN102722707A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1085456A2 (en) * | 1999-09-15 | 2001-03-21 | Siemens Corporate Research, Inc. | Character segmentation method for vehicle license plate recognition |
| CN101154271A (en) * | 2006-09-30 | 2008-04-02 | 电子科技大学中山学院 | License plate character segmentation method based on fast area labeling algorithm and license plate large-spacing locating method |
| CN102184399A (en) * | 2011-03-31 | 2011-09-14 | 上海名图信息技术有限公司 | Character segmenting method based on horizontal projection and connected domain analysis |
Non-Patent Citations (3)
| Title |
|---|
| 张全法等: "彩色图像的快速高保真灰度化方法研究", 《郑州大学学报(理学版)》, vol. 43, no. 3, 30 September 2011 (2011-09-30) * |
| 童立靖等: "几种文本图像二值化方法的对比分析", 《北方工业大学学报》, vol. 23, no. 01, 31 March 2011 (2011-03-31) * |
| 黄红燕等: "文字图像的整体倾斜矫正", 《计算机工程与应用》, vol. 42, no. 6, 15 March 2006 (2006-03-15) * |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103390160A (en) * | 2013-07-24 | 2013-11-13 | 佳都新太科技股份有限公司 | License plate character segmentation algorithm based on connected domain and template |
| CN104636748A (en) * | 2013-11-14 | 2015-05-20 | 张伟伟 | License plate recognition method and device |
| CN104200210B (en) * | 2014-08-12 | 2018-11-06 | 合肥工业大学 | A kind of registration number character dividing method based on component |
| CN104200210A (en) * | 2014-08-12 | 2014-12-10 | 合肥工业大学 | License plate character segmentation method based on parts |
| CN105678207A (en) * | 2014-11-19 | 2016-06-15 | 富士通株式会社 | Device and method for identifying content of target nameplate image from given image |
| 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 |
| CN106446898A (en) * | 2016-09-14 | 2017-02-22 | 宇龙计算机通信科技(深圳)有限公司 | Extraction method and extraction device of character information in image |
| CN106709394A (en) * | 2016-12-12 | 2017-05-24 | 北京慧眼智行科技有限公司 | Image processing method and device |
| CN106709394B (en) * | 2016-12-12 | 2019-07-05 | 北京慧眼智行科技有限公司 | A kind of image processing method and device |
| CN106650741A (en) * | 2016-12-30 | 2017-05-10 | 深圳市捷顺科技实业股份有限公司 | License plate character recognition method and device |
| CN109034157A (en) * | 2017-06-08 | 2018-12-18 | 北京君正集成电路股份有限公司 | Licence plate recognition method and device |
| CN107273892A (en) * | 2017-06-12 | 2017-10-20 | 北京智芯原动科技有限公司 | A kind of registration number character dividing method and device |
| CN107273892B (en) * | 2017-06-12 | 2020-06-16 | 北京智芯原动科技有限公司 | License plate character segmentation method and device |
| CN107392111A (en) * | 2017-06-27 | 2017-11-24 | 青岛海信电器股份有限公司 | Advertisement recognition method and device |
| CN107392111B (en) * | 2017-06-27 | 2020-06-23 | 海信视像科技股份有限公司 | Advertisement identification method and device |
| CN107610132A (en) * | 2017-08-28 | 2018-01-19 | 西北民族大学 | A kind of ancient books file and picture greasiness removal method |
| CN107610132B (en) * | 2017-08-28 | 2021-12-31 | 西北民族大学 | Method for removing stains from ancient book document image |
| CN107527056A (en) * | 2017-09-01 | 2017-12-29 | 南京邮电大学 | A kind of character segmentation method based on coarse positioning car plate |
| CN108256526A (en) * | 2017-12-07 | 2018-07-06 | 上海理工大学 | A kind of automobile license plate position finding and detection method based on machine vision |
| CN108256526B (en) * | 2017-12-07 | 2022-01-18 | 上海理工大学 | Motor vehicle license plate positioning detection method based on machine vision |
| CN109961553A (en) * | 2017-12-26 | 2019-07-02 | 航天信息股份有限公司 | Invoice number recognition methods, device and tax administration self-service terminal system |
| CN108734167A (en) * | 2018-05-08 | 2018-11-02 | 湖南开启时代电子信息技术有限公司 | A kind of contaminated film character recognition method |
| CN109086767A (en) * | 2018-07-11 | 2018-12-25 | 于洋 | The dividing method of fuzzy license plate in a kind of monitoring scene |
| CN109086767B (en) * | 2018-07-11 | 2019-05-07 | 于洋 | The dividing method of fuzzy license plate in a kind of monitoring scene |
| CN110728275B (en) * | 2018-07-16 | 2023-07-07 | 杭州海康威视数字技术股份有限公司 | License plate recognition method, license plate recognition device and storage medium |
| CN110728275A (en) * | 2018-07-16 | 2020-01-24 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device and storage medium |
| CN109241975A (en) * | 2018-08-27 | 2019-01-18 | 电子科技大学 | A kind of registration number character dividing method based on character center point location |
| CN109241975B (en) * | 2018-08-27 | 2022-03-15 | 电子科技大学 | License plate character segmentation method based on character center point positioning |
| CN109271988A (en) * | 2018-08-30 | 2019-01-25 | 中国传媒大学 | A kind of subtitle extraction method based on image segmentation and dynamic threshold |
| CN111563402A (en) * | 2019-02-14 | 2020-08-21 | 杭州海康威视数字技术股份有限公司 | License plate recognition method and device, terminal and storage medium |
| CN111563402B (en) * | 2019-02-14 | 2023-05-09 | 杭州海康威视数字技术股份有限公司 | License plate recognition method, license plate recognition device, terminal and storage medium |
| CN112836696A (en) * | 2019-11-22 | 2021-05-25 | 搜狗（杭州）智能科技有限公司 | Text data detection method and device and electronic equipment |
| CN112381088A (en) * | 2020-11-11 | 2021-02-19 | 上善智城(苏州)信息科技有限公司 | License plate recognition method and system for oil tank truck |
| CN113096095A (en) * | 2021-04-12 | 2021-07-09 | 江苏理工学院 | Machine vision-based adhesive tablet counting method and counting system |
| CN114549811A (en) * | 2021-12-31 | 2022-05-27 | 哈尔滨理工大学 | License plate identification method based on license plate spacer body characteristics |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102722707A (en) | License plate character segmentation method based on connected region and gap model | |
| CN105373794B (en) | A kind of licence plate recognition method | |
| CN101334836B (en) | License plate positioning method incorporating color, size and texture characteristic | |
| Alvarez et al. | Road detection based on illuminant invariance | |
| CN103116751B (en) | A kind of Method of Automatic Recognition for Character of Lcecse Plate | |
| CN103927526B (en) | Vehicle detecting method based on Gauss difference multi-scale edge fusion | |
| CN102509098B (en) | Fisheye image vehicle identification method | |
| CN101408942B (en) | Method for locating license plate under a complicated background | |
| CN103136528B (en) | A kind of licence plate recognition method based on dual edge detection | |
| CN109726717B (en) | Vehicle comprehensive information detection system | |
| CN101154271A (en) | License plate character segmentation method based on fast area labeling algorithm and license plate large-spacing locating method | |
| CN102096821B (en) | Number plate identification method under strong interference environment on basis of complex network theory | |
| CN103488978B (en) | License plate positioning method based on gray level jump and character projection interval mode | |
| CN102375982A (en) | Multi-character characteristic fused license plate positioning method | |
| CN100385452C (en) | Registration number character dividing method | |
| CN107016362B (en) | Vehicle weight recognition method and system based on vehicle front windshield pasted mark | |
| JP2005523530A (en) | System and method for identifying and extracting character string from captured image data | |
| CN101877050A (en) | Automatic extracting method for characters on license plate | |
| CN104732220A (en) | Specific color human body detection method oriented to surveillance videos | |
| CN109448000A (en) | A kind of dividing method of road sign image | |
| CN113837094A (en) | Road condition rapid analysis method based on full-color high-resolution remote sensing image | |
| CN110163109B (en) | Lane line marking method and device | |
| CN104200207A (en) | License plate recognition method based on Hidden Markov models | |
| CN111259796A (en) | Lane line detection method based on image geometric features | |
| CN104778833A (en) | Traffic light recognition 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 | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20121010 |