CN102496019A - License plate character segmenting method - Google Patents
License plate character segmenting method Download PDFInfo
- Publication number
- CN102496019A CN102496019A CN2011104052270A CN201110405227A CN102496019A CN 102496019 A CN102496019 A CN 102496019A CN 2011104052270 A CN2011104052270 A CN 2011104052270A CN 201110405227 A CN201110405227 A CN 201110405227A CN 102496019 A CN102496019 A CN 102496019A
- Authority
- CN
- China
- Prior art keywords
- zone
- license plate
- cut
- character
- height
- 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
Abstract
The invention discloses a license plate character segmenting method which comprises the following steps of: S10. correcting an inclined license plate; S11. removing an upper frame and a lower frame of the license plate; S12. segmenting license plate characters; and S13. normalizing the character size. The license plate character segmenting method overcomes the shortcomings that the traditional license plate character segmenting method is greatly affected by the background of the license plate, is high in the partitioning error rate, huge in calculation amount and long in execution time; the license plate character segmenting method is based on vertical projections and center-to-center spacing of the characters; and according to the license plate character segmenting method, firstly, a vertical projection method is used for realizing the coarse segmentation of the license plate characters, and then a first or second character segmentation region is found according to the center-to-center spacing of the license plate characters, the widths of the license plate characters and the prior knowledge of the license plate, and finally the merging of Chinese characters and the fine segmentation of numbers and English characters are realized. The license plate character segmenting method has the advantages that the computation complexity is low, the capacity of resisting disturbance is strong, the characters can be segmented from the background and the reliability is very high.
Description
Technical field
The invention belongs to the license plate recognition technology field of intelligent transportation aspect, specifically, relate to a kind of registration number character dividing method.
Background technology
Along with the continuous increase with automobile quantity that develops rapidly of various countries' highway construction, the task of traffic administration is heavy day by day, utilizes computing machine automotive license plate recognition technology, detects automatically with the identification automobile interior trade mark to have important effect in the traffic monitoring in modern times.Car plate detect with recognition technology be that Digital Image Processing and mode identification technology are at intelligent transportation (Intelligent Transportation System; ITS) one of important subject in the field; Its development to development of ITS and communications plays important impetus, has more vast market prospect.
As shown in Figure 1, car plate detects with recognition technology and is divided into four steps: car plate location, car plate correction, Character segmentation and character recognition.Characters on license plate cut apart be car plate detect with recognition technology in important component part, have only to accomplish effectively to cut apart could further extract the target character characteristic and discern.
Registration number character dividing method commonly used at present has sciagraphy, connected domain method and clustering methodology etc.Wherein, sciagraphy is present the most frequently used registration number character dividing method, and its thought is the characteristics according to character, carries out the license plate image after the binaryzation projection of vertical direction.Concentrate because the pixel of character more, and have certain gap to separate between each characters on license plate.The get off projected image that obtains of projection should have seven projection peak cluster of concentrating relatively like this, cuts apart the character that just can obtain car plate according to the minimum point between peak value then.Because the projection of character block in the vertical direction not only obtains local minimum at intercharacter; And local minimum also can be obtained in the gap in character; Therefore traditional projection dividing method is easy to be divided into two parts or three parts to Chinese character, causes segmentation errors.In addition, the spaced points between car plate left and right side frame and two three-character doctrines also all can interfere with projection to be cut apart, and causes segmentation errors; And to Character segmentation weak effect in the captured image under the different illumination conditions, poor anti jamming capability.
The connected domain method be with the object pixel on the image level line as starting point, extract the whole connected domains that comprise these initial points in the image through region growing.Write as because the letter and number in the car plate all is one, promptly include only a connected component, so each connected domain is a character, remainder in the image (part of not visiting in the area growth process) will be as noise remove.This method is very high for removing the noise requirement, because the phenomenon of character and the adhesion of car plate edge very general (particularly passing through the rivet at the second and the 6th character place), this will cause extracting a plurality of characters as a character, cause segmentation errors.In addition, many Chinese characters comprise a plurality of connected domains after binaryzation, and the digital and alphabetical phenomenon that the stroke fracture after binaryzation, can occur, therefore with the method meeting lost part character information, Character segmentation is easy to make a mistake.
Clustering methodology is to utilize in the pattern-recognition cluster algorithm to realize that characters on license plate cuts apart.It can solve the disconnected problem of Chinese character preferably, has solved the noise that exists during characters on license plate is cut apart preferably, and the car plate wearing and tearing cause problems such as character adhesion.In addition, it can handle some new-type car plates preferably through changing the coordinate of the preset type heart.But the programmed logic complex design of this method, loop nesting is more, and the processing time is long in real time.Simultaneously in order to improve the precision of the preset type heart, the width of car plate there is certain restriction.
It is thus clear that there are many defectives in the prior art registration number character dividing method: traditional sciagraphy and connected domain method receive the influence of car plate background bigger, and to receive when polluting its segmentation error rate very high when characters on license plate, thereby directly have influence on the Recognition of License Plate Characters of back; Though and the cluster analysis rule can overcome some complicated background slightly, but its calculated amount is big, to the rate request of computing machine than higher; Execution time is long, and is more consuming time, and for broadside apart from license plate image, easily back gauge is divided into character, cause erroneous judgement.
So, be necessary to study in fact, a kind of good real time performance that has is provided, computation complexity is low, can from complicated car plate background, accurately be partitioned into the technical scheme of character.
Summary of the invention
For addressing the above problem, the object of the present invention is to provide a kind of good real time performance that has, computation complexity is low, can from complicated car plate background, accurately be partitioned into the registration number character dividing method of character.
For realizing above-mentioned purpose, technical scheme of the present invention is:
A kind of registration number character dividing method comprises the steps:
S10: license plate sloped correction;
S11: the car plate upper and lower side frame is removed;
S12: characters on license plate is cut apart;
S13: character boundary normalization.
Further, said step S10 includes following steps:
S100: carry out the Canny rim detection to locating the original gray scale license plate image that obtains;
S101: to accomplishing the license plate image probability of use Hough change detection straight line of Canny rim detection, and calculate the angle of inclination of car plate;
S102: according to the resulting license plate sloped angle of step S101 original gray scale license plate image is carried out the rotation of respective angles, to obtain horizontal gray scale license plate image.
Further, said step S11 includes following steps:
S110: use the Otsu threshold method to carry out binaryzation to the horizontal gray scale license plate image that obtains among the step S102, obtain the two-value license plate image;
S111: utilize the Gray Level Jump method to remove the car plate upper and lower side frame of two-value license plate image.
Further, said step S12 includes following steps:
S120: carry out carrying out rough segmentation again after the vertical projection and cut to removing two-value license plate image behind the car plate upper and lower side frame;
S121: utilize license plate image after character center spacing and car plate priori are cut rough segmentation to segment and cut.
Registration number character dividing method of the present invention has overcome existing registration number character dividing method and has received the deficiency that the car plate background influence is big, segmentation error rate is high, calculated amount is big, the execution time is long; Its characters on license plate based on vertical projection and character center spacing is cut apart; Utilize the rough segmentation of vertical projection method's realization characters on license plate to cut earlier; Find first or second Character segmentation zone of car plate then according to character center spacing, characters on license plate width and the car plate priori of car plate, realize that the merging of Chinese character and the segmentation of numeral and English character cut.Registration number character dividing method computation complexity of the present invention is low, and antijamming capability is strong, can from complex background, be partitioned into character, has very high reliability.
Description of drawings
Fig. 1 is that car plate detects and identification process figure in the prior art.
Fig. 2 is a Character segmentation process flow diagram of the present invention.
Fig. 3 is the original gray scale license plate image of example with 7 characters for the present invention.
Fig. 4 is the license plate image after Fig. 3 Canny rim detection.
The straight line that Fig. 5 arrives for Fig. 3 probability Hough change detection.
Fig. 6 is the license plate image after Fig. 3 slant correction.
Fig. 7 is the license plate image after Fig. 3 Otsu method binaryzation.
Fig. 8 removes upper and lower side frame license plate image afterwards for Fig. 3.
Fig. 9 is the license plate image after Fig. 3 Character segmentation.
Figure 10 split for Fig. 3 and normalization after characters on license plate diagram.
Embodiment
In order to make the object of the invention, technical scheme and advantage clearer,, the present invention is further elaborated below in conjunction with accompanying drawing and embodiment.Should be appreciated that specific embodiment described herein only in order to explanation the present invention, and be not used in qualification the present invention.
Please with reference to shown in Figure 2, registration number character dividing method of the present invention may further comprise the steps:
S10: license plate sloped correction;
S11: the car plate upper and lower side frame is removed;
S12: characters on license plate is cut apart;
S13: character boundary normalization.
Wherein,
The license plate sloped correction of step S10 includes following steps:
S100: carry out the Canny rim detection to locating the original gray scale license plate image that obtains;
S101: to accomplishing the license plate image probability of use Hough change detection straight line of Canny rim detection, and calculate the angle of inclination of car plate;
S102: according to the resulting license plate sloped angle of step S101 original gray scale license plate image is carried out the rotation of respective angles, to obtain horizontal gray scale license plate image.
Step S11 car plate upper and lower side frame is removed and is included following steps:
S110: use the Otsu threshold method to carry out binaryzation to the horizontal gray scale license plate image that obtains among the step S102, obtain the two-value license plate image;
S111: utilize the Gray Level Jump method to remove the car plate upper and lower side frame of two-value license plate image.
Step S12 characters on license plate is cut apart and is included following steps:
S120: carry out carrying out rough segmentation again after the vertical projection and cut to removing two-value license plate image behind the car plate upper and lower side frame;
S121: utilize license plate image after character center spacing, characters on license plate width and car plate priori are cut rough segmentation to segment and cut.Wherein, the car plate priori comprises that mainly the ratio of width to height of actual license plate character and first and second character of car plate do not comprise character " 1 ".
Please extremely shown in Figure 10 with reference to Fig. 3, the embodiment of the invention is that example describes with car plate capital LR8251.
S10: license plate sloped correction
In the ideal case, license plate image should be a rectangle, receives camera lens and licence plate angle during still owing to shooting, vehicle movement, and the influence of pavement behavior etc., license plate inclination to a certain degree occurs through regular meeting, also might produce distortion.The car plate that tilts can then have influence on the vertical projection of car plate, and then causes the Character segmentation mistake indirectly.Therefore, be necessary the license plate image that tilts is proofreaied and correct.License plate sloped correction body is specific as follows:
S100: carry out the Canny rim detection to locating the original gray scale license plate image that obtains.The angle of inclination of car plate is to calculate according to the degree of tilt of car plate upper and lower side frame, thus use the edge of the outstanding original gray scale license plate image of Canny edge detection method here, so that next step detects the frame straight line.Used the computer vision storehouse OpenCV that increases income of Intel Company in the present embodiment, the function prototype of Canny rim detection is following:
void?cvCanny(const?CvArr*image，CvArr*edges，double?threshold1，double?threshold2，int?aperture_size＝3)；
Image representing input images wherein, edges represents output image, and the little threshold value in the middle of threshold1 and the threshold is used for controlling edge link, and big threshold value is used for controlling the initial segmentation at edge, and aperture_size is a Sobel operator kernel size.The function and the parameter that adopt in the present embodiment are following:
cvCanny(image，edges，50，200，3)；
The effect of license plate image is as shown in Figure 4 after the completion Canny rim detection.
S101: to accomplishing the license plate image probability of use Hough change detection straight line of Canny rim detection, and calculate the angle of inclination of car plate.Straight-line detection is normally utilized the prior Hough conversion algorithm, but this algorithm need scan calculating to each pixel of entire image, and calculated amount is big, and the processing time is long; Adopted probability Hough conversion to come the straight line in the detected image among the present invention, to reduce computing time.The function prototype of probability Hough conversion in OpenCV is following:
CvSeq*cvHoughLines2(CvArr*image，void*line_storage，int?method，double?rho，double?theta，int?threshold，double?param1＝0，double?param2＝0)；
Wherein, image is an input picture, and line_storage is detected line segment storage silo; Method is the Hough transformed variable; Rho is and the range accuracy of pixel relevant unit, and theta is the angle precision of arc measurement, and threshold is a threshold parameter; Param1 is a minimum line segment length, and param2 is illustrated in and carries out the largest interval that broken line segment connects on same the straight line.The function and the parameter that adopt in the present embodiment are following:
cvHoughLines2(image，lines_storage，CV_HOUGH_PROBABILISTIC，1，CV_PI/180，80，30，10)；
Adopting after this function detects line segment, calculating the slope and the angle of inclination of each line segment, and from less than selecting the angle of inclination of maximum angle as car plate the angle of inclination of 30 degree.The line segment that probability Hough change detection arrives is as shown in Figure 5.
S102: to original gray scale license plate image, carry out the rotation of respective angles according to detected angle of inclination, obtain horizontal gray scale license plate image.Adopt the bilinear interpolation algorithm that original gray scale license plate image is rotated in the present embodiment, can keep the details in the image rotating, its postrotational effect such as Fig. 6, visible rotation back characters on license plate does not deform.
S11: the car plate upper and lower side frame is removed
Also can there be the interference of upper and lower side frame in car plate through behind the slant correction, if do not remove, and will be to character cutting and identification deleterious impact.The removal of car plate upper and lower side frame is specific as follows:
S110: use the Otsu threshold method to carry out binaryzation to the horizontal gray scale license plate image that obtains, obtain the two-value license plate image.In the present embodiment, press the gamma characteristic of image, image is divided into background and prospect two parts.The Otsu threshold method obtains segmentation threshold through calculating maximum between-cluster variance, and " under the bimodal condition " comparatively desirable can access segmentation effect preferably with the Otsu threshold method.Binaryzation function prototype among the OpenCV is following:
void?cvThreshold(const?CvArr*src，CvArr*dst，double?threshold，doublemax_value，int?threshold_type)；
Wherein, src is original array, and dst is the output array, and threshold is a threshold value, and max_value is for using the maximal value of CV_THRESH_BINARY and CV_THRESH_BINARY_INV, and threshold_type is a threshold type.Function and the parameter used in the present embodiment are following:
cvThreshold(src，dst，100，255，CV_THRESH_OTSU)；
As shown in Figure 7, it is to utilize the Otsu threshold method gray scale license plate image through overcorrect to be carried out the bianry image that obtains after the binaryzation.
S111: utilize the Gray Level Jump method to remove the car plate upper and lower side frame of two-value license plate image.According to the car plate priori, through on the car plate after the binaryzation 7 characters being arranged, the saltus step of each character is counted and is at least 2, and at character zone, its horizontal black and white trip point sum is greater than 14, but not character zone does not then satisfy this characteristic.So can utilize black and white saltus step method to remove the car plate upper and lower side frame, it comprises the steps:
S1110: begin upwards scanning from 1/2nd of bianry image, the black and white number of transitions of adding up each row is if its value explains that then this row has not been a character zone, its first trip as character zone less than 14;
S1111: begin downward scanning from 1/2nd of bianry image, add up the black and white number of transitions of each row, if its value explains that then this row has not been a character zone, its footline as character zone less than 14;
S1112: the zone between first trip and the footline is character zone, has so just removed the influence of upper and lower side frame.
As shown in Figure 8, it shows the design sketch of removing license plate image after the frame computing.
S12: characters on license plate is cut apart
S120: carry out carrying out rough segmentation again after the vertical projection and cut to removing two-value license plate image behind the car plate upper and lower side frame.Wherein, concrete steps are following:
S1200: from removing the left side begin column scanning of the two-value license plate image behind the car plate upper and lower side frame, when the black picture element number less than certain threshold value T (getting 2 in the present embodiment), then note these row a, put into array ColNo [2i];
S1201:, put into array ColNo [2i+1] if the black picture element number is then noted these row b greater than certain threshold value (getting 2 in the present embodiment); The a that integrating step S1200 obtains just can confirm first cut zone, and a and b are respectively the left and right border of cut zone;
S1202: continue scanning according to above-mentioned steps S1200, step S1201 and go down, after having scanned all row, what deposited the even number position among the array ColNo [i] is the left margin of cut zone, and what odd positions was deposited is the right margin of cut zone.
S121: utilize license plate image after character center spacing and car plate priori are cut rough segmentation to segment and cut.Actual license plate height with domestic common car is an example, and the actual license plate height of domestic common car is 140mm, and width is 440mm; The actual characters height is 90mm, and except that " 1 ", the developed width of other single characters is 45mm; The actual characters center distance is 57mm (two three-character doctrine center distance is 79mm).Therefore the wide high ratio of actual characters is 1: 2, and the ratio of actual characters center distance and characters on license plate height is 57: 90 and 79: 90 (ratio of two three-character doctrine center distance and characters on license plate height).So under the situation of knowing the characters on license plate height, just can confirm the width range of character and the scope of character center spacing according to the ratio of actual license plate the ratio of width to height and character center spacing and characters on license plate height.Utilize above character center spacing and car plate priori just can realize that the segmentation of characters on license plate cuts, it specifically comprises the steps:
S1210: the region quantity that the inspection rough segmentation cuts out.If quantity is less than 7 then get into subsequent step S1211; If quantity greater than 15 then cut apart failure, quits a program; If in the middle of 7 and 15 then get into subsequent step S1212.
S1211: carry out a rough segmentation again after the increase threshold value and cut, and inspection area quantity.If quantity is less than 7 or greater than 15 then cut apart failure, quit a program.
S1212: check the width of each cut zone, compare with setting threshold MaxThresholdWidth (this threshold value can be confirmed according to the car plate height (height) that navigates to, be taken as height/2+height/10+2.5 in the present embodiment).If greater than this threshold value then increase once more and image is carried out a rough segmentation again after the threshold value and cut; Do not cut if carry out rough segmentation, then directly get into subsequent step S1214.
S1213: the region quantity that inspection splits.If greater than 15, then cut apart failure, quit a program.
S1214: the width of checking each cut zone.If the width that also has the zone is then cut apart failure greater than MaxThresholdWidth, quit a program.
S1215: add up the white pixel number of spots in each cut zone.If cut zone quantity is less than 7, then cut apart failure, quit a program; If the white pixel number of spots in the cut zone is then abandoned this zone less than a certain threshold value ThresholdNo (being taken as height*2 in the present embodiment).Thereby can filter out some little noise spots and second and three-character doctrine between that separation.
S1216: the width of checking each cut zone.If its width is less than a certain threshold value (being made as height/10-0.5 here), then remove it.This operation can be removed some and disturb cut zone, but guarantees can not remove character " 1 ".
S1217: inspection cut zone quantity: if be less than 7, then cut apart failure, quit a program.
S1218: search for next peak width satisfy certain limit (be taken as the cut zone Ω of (height/2-height/10-1.5, height/2+height/10+1.5)) here, if the search less than, then cut apart failure, quit a program.
S1219: the center of inspection cut zone Ω and the horizontal range of next cut zone center, check whether it equals T1=car plate height * 57/90 or equal T2=car plate height * 79/90.If equal T1, then regional as first characters on license plate cut zone Ω, and get into step S1220; If equal T2, then cut zone Ω as second character zone of car plate, get into step S1221; If all do not meet with T1 and T2, then search for the cut zone Ω that next peak width satisfies certain limit, repeated execution of steps S1219.
S1220: the cut zone Ω that finds is first characters on license plate zone (Chinese character zone), judges the cut zone quantity of its back, has promptly formed the characters on license plate zone if quantity, then reads next six zones in proper order more than or equal to 6, cuts apart end, quits a program; If quantity less than 6, is then cut apart failure, quit a program.
S1221: the cut zone Ω that finds is second character zone, explains that Chinese character possibly be made up of two parts or three parts, then checks the cut zone quantity of back, if be less than 5, then cuts apart failure, quits a program.
S1222: the cut zone quantity of Ω front, inspection area.If less than 2, then cut apart failure, quit a program; If equal 2, get into step S1223; If more than or equal to 3, get into step S1224.
S1223: merge two zones, regional Ω front, and the peak width after the inspection merging.If satisfying certain limit, its width (is taken as (height/2-height/10-1.5 in the present embodiment; Height/2+height/10+1.5)); And the central horizontal distance that merges zone and regional Ω is roughly car plate height * 57/90, thinks that then the zone after merging is the Chinese character zone.Then order reads five zones of regional Ω back, forms the characters on license plate zone, cuts apart end, quits a program; If failure is then cut apart in one of them requirement above not satisfying, quit a program.
S1224: merge two zones, regional Ω front earlier, and the peak width after the inspection merging.If satisfying certain limit, its width (is taken as (height/2-height/10-1.5 in the present embodiment; Height/2+height/10+1.5)); And the central horizontal distance that merges zone and regional Ω is roughly car plate height * 57/90, thinks that then the zone after merging is the Chinese character zone.Then order reads five zones of regional Ω back, forms the characters on license plate zone, cuts apart end, quits a program; If the width after merging is then cut apart failure greater than (height/2+height/10+1.5), quit a program; If the width after merging then gets into step S1225 less than (height/2-height/10-1.5).
S1225: merge three zones, regional Ω front, and the peak width after the inspection merging.If satisfying certain limit, its width (is taken as (height/2-height/10-1.5 in the present embodiment; Height/2+height/10+1.5)); And the central horizontal distance that merges zone and regional Ω is roughly car plate height * 57/90, thinks that then the zone after merging is the Chinese character zone.Then order reads five zones of regional Ω back, forms the characters on license plate zone, cuts apart end, quits a program; If failure is then cut apart in one of them requirement above not satisfying, quit a program.
As shown in Figure 9, it shows the license plate image after the Character segmentation.
S13: character boundary normalization
The character that license plate image not of uniform size often causes splitting is also not of uniform size, in order to help follow-up character recognition, also need carry out normalization to the character boundary that cuts down.Normalize to the 36x20 size to character in the present embodiment.In OpenCV, the function prototype that changes the image size is following:
Void?cvResize(const?CvArr*src，CvArr*dst，int?interpolation＝CV_INTER_LINEAR)；
Src is an input picture, and dst is an output image, and interpolation is an interpolation method.Function and the parameter used in the present embodiment are following:
cvResize(src，dst，CV_INTER_NN)。
Shown in figure 10, its show split and normalization after characters on license plate diagram.
The above is merely preferred embodiment of the present invention, not in order to restriction the present invention, all any modifications of within spirit of the present invention and principle, being done, is equal to and replaces and improvement etc., all should be included within protection scope of the present invention.
Claims (9)
1. a registration number character dividing method is characterized in that, comprises the steps:
S10: license plate sloped correction;
S11: the car plate upper and lower side frame is removed;
S12: characters on license plate is cut apart;
S13: character boundary normalization.
2. registration number character dividing method as claimed in claim 1 is characterized in that: said step S10 includes following steps:
S100: carry out the Canny rim detection to locating the original gray scale license plate image that obtains;
S101: to accomplishing the license plate image probability of use Hough change detection straight line of Canny rim detection, and calculate the angle of inclination of car plate;
S102: according to the resulting license plate sloped angle of step S101 original gray scale license plate image is carried out the rotation of respective angles, to obtain horizontal gray scale license plate image.
3. registration number character dividing method as claimed in claim 2 is characterized in that: said step S11 includes following steps:
S110: use the Otsu threshold method to carry out binaryzation to the horizontal gray scale license plate image that obtains among the step S102, obtain the two-value license plate image;
S111: utilize the Gray Level Jump method to remove the car plate upper and lower side frame of two-value license plate image.
4. registration number character dividing method as claimed in claim 3 is characterized in that: said step S12 includes following steps:
S120: carry out carrying out rough segmentation again after the vertical projection and cut to removing two-value license plate image behind the car plate upper and lower side frame;
S121: utilize license plate image after character center spacing and car plate priori are cut rough segmentation to segment and cut.
5. registration number character dividing method as claimed in claim 4 is characterized in that: in said step S102, adopt the bilinear interpolation algorithm that original gray scale license plate image is rotated.
6. registration number character dividing method as claimed in claim 5 is characterized in that: in said step S110, press the gamma characteristic of image, use the Otsu algorithm that image is divided into background and prospect two parts.
7. registration number character dividing method as claimed in claim 6 is characterized in that: said step S111 comprises the steps:
S1110: begin upwards scanning from 1/2nd of bianry image, the black and white number of transitions of adding up each row is if its value explains that then this row has not been a character zone, its first trip as character zone less than 14;
S1111: begin downward scanning from 1/2nd of bianry image, add up the black and white number of transitions of each row, if its value explains that then this row has not been a character zone, its footline as character zone less than 14;
S1112: the zone between first trip and the footline is character zone, has so just removed the influence of upper and lower side frame.
8. registration number character dividing method as claimed in claim 7 is characterized in that: said step S120 comprises the steps:
S1200: from removing the left side begin column scanning of the two-value license plate image behind the car plate upper and lower side frame,, then note these row a, put into array ColNo [2i] when black picture element number during less than certain threshold value T;
S1201:, put into array ColNo [2i+1] if the black picture element number, is then noted these row b greater than certain threshold value; The a that integrating step S1200 obtains just can confirm first cut zone, and a and b are respectively the left and right border of cut zone;
S1202: continue scanning according to above-mentioned steps S1200, step S1201 and go down, after having scanned all row, what deposited the even number position among the array ColNo [i] is the left margin of cut zone, and what odd positions was deposited is the right margin of cut zone.
9. registration number character dividing method as claimed in claim 8 is characterized in that: said step S121 comprises the steps:
S1210: the inspection rough segmentation region quantity that cuts out, if quantity is less than 7 then get into subsequent step S1211; If quantity greater than 15 then cut apart failure, quits a program; If in the middle of 7 and 15 then get into subsequent step S1212;
S1211: increase and to carry out a rough segmentation again after the threshold value and cut, and inspection area quantity,, quit a program if quantity is less than 7 or greater than 15 then cut apart failure;
S1212: check the width of each cut zone, with setting threshold MaxThresholdWidth relatively, if greater than this threshold value then increase once more and image is carried out a rough segmentation again after the threshold value and cut; Do not cut if carry out rough segmentation, then directly get into subsequent step S1214.
S1213: the region quantity that inspection splits, if greater than 15, then cut apart failure, quit a program;
S1214: check the width of each cut zone,, quit a program if the width that also has the zone is then cut apart failure greater than MaxThresholdWidth;
S1215: add up the white pixel number of spots in each cut zone; If the white pixel number of spots in the cut zone is less than a certain threshold value ThresholdNo; Then abandon this zone, thus can filter out some little noise spots and second and three-character doctrine between that separation;
S1216: check the width of each cut zone, if its width less than a certain threshold value, is then removed it, this operation can be removed some and disturb cut zone, but guarantees can not remove character " 1 ";
S1217: inspection cut zone quantity: if be less than 7, then cut apart failure, quit a program;
S1218: search for the cut zone Ω of next peak width between height/2-height/10-1.5 and height/2+height/10+1.5, if the search less than, then cut apart failure, quit a program;
S1219: the center of inspection cut zone Ω and the horizontal range of next cut zone center, check whether it equals T1=car plate height * 57/90 or equal T2=car plate height * 79/90; If equal T1, then regional as first characters on license plate cut zone Ω, and get into step S1220; If equal T2, then cut zone Ω as second character zone of car plate, get into step S1221; If all do not meet with T1 and T2, then search for the cut zone Ω that next peak width satisfies certain limit, repeated execution of steps S1219;
S1220: the cut zone Ω that finds is first characters on license plate zone, judges the cut zone quantity of its back, has promptly formed the characters on license plate zone if quantity, then reads next six zones in proper order more than or equal to 6, cuts apart end, quits a program; If quantity less than 6, is then cut apart failure, quit a program;
S1221: the cut zone Ω that finds is second character zone, explains that Chinese character possibly be made up of two parts or three parts, then checks the cut zone quantity of back, if be less than 5, then cuts apart failure, quits a program;
S1222: the cut zone quantity of Ω front, inspection area, if less than 2, then cut apart failure, quit a program; If equal 2, get into step S1223; If more than or equal to 3, get into step S1224;
S1223: merge two zones, regional Ω front, and the peak width after the inspection merging, if its width is between height/2-height/10-1.5 and height/2+height/10+1.5; And the central horizontal distance that merges zone and regional Ω is roughly car plate height * 57/90; Think that then the zone after merging is the Chinese character zone, then order reads five zones of regional Ω back, forms the characters on license plate zone; Cut apart end, quit a program; If failure is then cut apart in one of them requirement above not satisfying, quit a program;
S1224: merge two zones, regional Ω front earlier, and the peak width after the inspection merging, if its width is between height/2-height/10-1.5 and height/2+height/10+1.5; And the central horizontal distance that merges zone and regional Ω is roughly car plate height * 57/90; Think that then the zone after merging is the Chinese character zone, then order reads five zones of regional Ω back, forms the characters on license plate zone; Cut apart end, quit a program; If the width after merging is then cut apart failure greater than height/2+height/10+1.5, quit a program; If the width after merging then gets into step S1225 less than height/2-height/10-1.5;
S1225: merge three zones, regional Ω front, and the peak width after the inspection merging; If its width is between height/2-height/10-1.5 and height/2+height/10+1.5, and the central horizontal that merges zone and regional Ω thinks then that apart from being roughly car plate height * 57/90 zone after merging is the Chinese character zone; Then order reads five zones of regional Ω back, forms the characters on license plate zone, cuts apart end, quits a program; If failure is then cut apart in one of them requirement above not satisfying, quit a program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011104052270A CN102496019A (en) | 2011-12-08 | 2011-12-08 | License plate character segmenting method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2011104052270A CN102496019A (en) | 2011-12-08 | 2011-12-08 | License plate character segmenting method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102496019A true CN102496019A (en) | 2012-06-13 |
Family
ID=46187844
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2011104052270A Pending CN102496019A (en) | 2011-12-08 | 2011-12-08 | License plate character segmenting method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102496019A (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103077388A (en) * | 2012-10-31 | 2013-05-01 | 浙江大学 | Rapid text scanning method oriented to portable computing equipment |
| CN103198315A (en) * | 2013-04-17 | 2013-07-10 | 南京理工大学 | License plate character segmentation algorithm based on character outline and template matching |
| CN103455815A (en) * | 2013-08-27 | 2013-12-18 | 电子科技大学 | Self-adaptive license plate character segmentation method in complex scene |
| CN104657726A (en) * | 2015-03-18 | 2015-05-27 | 哈尔滨工程大学 | Identification method for truck numbers of railway tank trucks |
| CN104700432A (en) * | 2015-03-24 | 2015-06-10 | 银江股份有限公司 | Self-adaptive adhered vehicle separating method |
| CN104834930A (en) * | 2015-05-25 | 2015-08-12 | 山东中孚信息产业股份有限公司 | Image Chinese character search method for multi system platform |
| CN105117727A (en) * | 2015-09-18 | 2015-12-02 | 浙江工业大学 | Method for fast positioning license plate |
| CN105608455A (en) * | 2015-12-18 | 2016-05-25 | 浙江宇视科技有限公司 | License plate tilt correction method and apparatus |
| CN105631425A (en) * | 2015-12-29 | 2016-06-01 | 厦门科拓通讯技术股份有限公司 | License plate recognition method based on video streams and system thereof and intelligent digital video camera |
| CN106529540A (en) * | 2016-11-07 | 2017-03-22 | 湖南源信光电科技有限公司 | Gray projection-based license plate character segmentation method |
| CN106529534A (en) * | 2016-11-07 | 2017-03-22 | 湖南源信光电科技有限公司 | Variable-length license plate character segmentation method based on hybrid tilt correction and projection method |
| CN106951896A (en) * | 2017-02-22 | 2017-07-14 | 武汉黄丫智能科技发展有限公司 | A kind of license plate image sloped correcting method |
| CN106980857A (en) * | 2017-02-24 | 2017-07-25 | 浙江工业大学 | A kind of Brush calligraphy segmentation recognition method based on rubbings |
| CN107992785A (en) * | 2016-10-26 | 2018-05-04 | 北京君正集成电路股份有限公司 | The recognition methods of fuzzy license plate and device |
| CN108446702A (en) * | 2018-03-14 | 2018-08-24 | 深圳怡化电脑股份有限公司 | A kind of image character dividing method, device, equipment and storage medium |
| CN108470175A (en) * | 2018-03-27 | 2018-08-31 | 康体佳智能科技（深圳）有限公司 | Vehicle License Plate Recognition System based on neural network and recognition methods |
| CN109034145A (en) * | 2018-07-06 | 2018-12-18 | 西安电子科技大学 | Bank's card number recognition methods based on OpenCV |
| CN109492562A (en) * | 2018-10-30 | 2019-03-19 | 河南科技大学 | A kind of contactless domestic animal identifying system |
| CN110059695A (en) * | 2019-04-23 | 2019-07-26 | 厦门商集网络科技有限责任公司 | A kind of character segmentation method and terminal based on upright projection |
| CN110335286A (en) * | 2019-07-11 | 2019-10-15 | 京东方科技集团股份有限公司 | Acquisition methods, computer equipment and the computer-readable medium of gray threshold |
| CN111027546A (en) * | 2019-12-05 | 2020-04-17 | 北京嘉楠捷思信息技术有限公司 | Character segmentation method and device and computer readable storage medium |
| CN111223065A (en) * | 2020-01-13 | 2020-06-02 | 中国科学院重庆绿色智能技术研究院 | Image correction method, irregular text recognition device, storage medium and equipment |
| CN111598104A (en) * | 2020-06-30 | 2020-08-28 | 成都鹏业软件股份有限公司 | License plate character recognition method and system |
| CN112215852A (en) * | 2020-09-29 | 2021-01-12 | 忻州师范学院 | Digital image segmentation method based on cluster learning device integration |
| US11615634B2 (en) | 2021-07-01 | 2023-03-28 | International Business Machines Corporation | Character recognition of license plate under complex background |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020159636A1 (en) * | 2000-03-14 | 2002-10-31 | Lienhart Rainer W | Generalized text localization in images |
| 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 |
-
2011
- 2011-12-08 CN CN2011104052270A patent/CN102496019A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20020159636A1 (en) * | 2000-03-14 | 2002-10-31 | Lienhart Rainer W | Generalized text localization in images |
| 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 |
Cited By (43)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103077388B (en) * | 2012-10-31 | 2016-01-20 | 浙江大学 | Fast text towards portable computing device sweeps the method for reading |
| CN103077388A (en) * | 2012-10-31 | 2013-05-01 | 浙江大学 | Rapid text scanning method oriented to portable computing equipment |
| CN103198315A (en) * | 2013-04-17 | 2013-07-10 | 南京理工大学 | License plate character segmentation algorithm based on character outline and template matching |
| CN103198315B (en) * | 2013-04-17 | 2015-12-02 | 南京理工大学 | Based on the Character Segmentation of License Plate of character outline and template matches |
| CN103455815B (en) * | 2013-08-27 | 2016-08-17 | 电子科技大学 | A kind of self adaptation registration number character dividing method under complex scene |
| CN103455815A (en) * | 2013-08-27 | 2013-12-18 | 电子科技大学 | Self-adaptive license plate character segmentation method in complex scene |
| CN104657726A (en) * | 2015-03-18 | 2015-05-27 | 哈尔滨工程大学 | Identification method for truck numbers of railway tank trucks |
| CN104700432A (en) * | 2015-03-24 | 2015-06-10 | 银江股份有限公司 | Self-adaptive adhered vehicle separating method |
| CN104700432B (en) * | 2015-03-24 | 2017-11-03 | 银江股份有限公司 | A kind of adaptive adhesion Method of Vehicle Segmentation |
| CN104834930A (en) * | 2015-05-25 | 2015-08-12 | 山东中孚信息产业股份有限公司 | Image Chinese character search method for multi system platform |
| CN104834930B (en) * | 2015-05-25 | 2018-07-24 | 中孚信息股份有限公司 | A kind of image method for looking up Chinese characters for multisystem platform |
| CN105117727A (en) * | 2015-09-18 | 2015-12-02 | 浙江工业大学 | Method for fast positioning license plate |
| CN105117727B (en) * | 2015-09-18 | 2018-05-08 | 浙江工业大学 | A kind of Location Method of Vehicle License Plate |
| CN105608455B (en) * | 2015-12-18 | 2019-06-11 | 浙江宇视科技有限公司 | A kind of license plate sloped correcting method and device |
| CN105608455A (en) * | 2015-12-18 | 2016-05-25 | 浙江宇视科技有限公司 | License plate tilt correction method and apparatus |
| CN105631425A (en) * | 2015-12-29 | 2016-06-01 | 厦门科拓通讯技术股份有限公司 | License plate recognition method based on video streams and system thereof and intelligent digital video camera |
| CN105631425B (en) * | 2015-12-29 | 2020-04-07 | 厦门科拓通讯技术股份有限公司 | License plate recognition method and system based on video stream and intelligent digital camera |
| CN107992785B (en) * | 2016-10-26 | 2021-08-31 | 北京君正集成电路股份有限公司 | Fuzzy license plate recognition method and device |
| CN107992785A (en) * | 2016-10-26 | 2018-05-04 | 北京君正集成电路股份有限公司 | The recognition methods of fuzzy license plate and device |
| CN106529534A (en) * | 2016-11-07 | 2017-03-22 | 湖南源信光电科技有限公司 | Variable-length license plate character segmentation method based on hybrid tilt correction and projection method |
| CN106529540A (en) * | 2016-11-07 | 2017-03-22 | 湖南源信光电科技有限公司 | Gray projection-based license plate character segmentation method |
| CN106951896A (en) * | 2017-02-22 | 2017-07-14 | 武汉黄丫智能科技发展有限公司 | A kind of license plate image sloped correcting method |
| CN106951896B (en) * | 2017-02-22 | 2020-01-03 | 武汉黄丫智能科技发展有限公司 | License plate image tilt correction method |
| CN106980857B (en) * | 2017-02-24 | 2020-05-05 | 浙江工业大学 | Chinese calligraphy segmentation and recognition method based on copybook |
| CN106980857A (en) * | 2017-02-24 | 2017-07-25 | 浙江工业大学 | A kind of Brush calligraphy segmentation recognition method based on rubbings |
| CN108446702A (en) * | 2018-03-14 | 2018-08-24 | 深圳怡化电脑股份有限公司 | A kind of image character dividing method, device, equipment and storage medium |
| CN108446702B (en) * | 2018-03-14 | 2022-05-31 | 深圳怡化电脑股份有限公司 | Image character segmentation method, device, equipment and storage medium |
| CN108470175A (en) * | 2018-03-27 | 2018-08-31 | 康体佳智能科技（深圳）有限公司 | Vehicle License Plate Recognition System based on neural network and recognition methods |
| CN109034145A (en) * | 2018-07-06 | 2018-12-18 | 西安电子科技大学 | Bank's card number recognition methods based on OpenCV |
| CN109034145B (en) * | 2018-07-06 | 2021-11-09 | 西安电子科技大学 | Bank card number identification method based on OpenCV |
| CN109492562B (en) * | 2018-10-30 | 2021-05-25 | 河南科技大学 | Non-contact livestock identification system |
| CN109492562A (en) * | 2018-10-30 | 2019-03-19 | 河南科技大学 | A kind of contactless domestic animal identifying system |
| CN110059695B (en) * | 2019-04-23 | 2021-08-27 | 厦门商集网络科技有限责任公司 | Character segmentation method based on vertical projection and terminal |
| CN110059695A (en) * | 2019-04-23 | 2019-07-26 | 厦门商集网络科技有限责任公司 | A kind of character segmentation method and terminal based on upright projection |
| CN110335286A (en) * | 2019-07-11 | 2019-10-15 | 京东方科技集团股份有限公司 | Acquisition methods, computer equipment and the computer-readable medium of gray threshold |
| CN111027546A (en) * | 2019-12-05 | 2020-04-17 | 北京嘉楠捷思信息技术有限公司 | Character segmentation method and device and computer readable storage medium |
| CN111027546B (en) * | 2019-12-05 | 2024-03-26 | 嘉楠明芯(北京)科技有限公司 | Character segmentation method, device and computer readable storage medium |
| CN111223065A (en) * | 2020-01-13 | 2020-06-02 | 中国科学院重庆绿色智能技术研究院 | Image correction method, irregular text recognition device, storage medium and equipment |
| CN111223065B (en) * | 2020-01-13 | 2023-08-01 | 中国科学院重庆绿色智能技术研究院 | Image correction method, irregular text recognition device, storage medium and apparatus |
| CN111598104A (en) * | 2020-06-30 | 2020-08-28 | 成都鹏业软件股份有限公司 | License plate character recognition method and system |
| CN111598104B (en) * | 2020-06-30 | 2023-05-12 | 成都鹏业软件股份有限公司 | License plate character recognition method and system |
| CN112215852A (en) * | 2020-09-29 | 2021-01-12 | 忻州师范学院 | Digital image segmentation method based on cluster learning device integration |
| US11615634B2 (en) | 2021-07-01 | 2023-03-28 | International Business Machines Corporation | Character recognition of license plate under complex background |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102496019A (en) | License plate character segmenting method | |
| US11062155B2 (en) | Monocular cued detection of three-dimensional structures from depth images | |
| US10943131B2 (en) | Image based lane marking classification | |
| US10521676B2 (en) | Lane detection device, lane departure determination device, lane detection method and lane departure determination method | |
| US7545956B2 (en) | Single camera system and method for range and lateral position measurement of a preceding vehicle | |
| CN102043959B (en) | License plate character segmentation method | |
| CN109583280A (en) | Lane detection method, apparatus, equipment and storage medium | |
| US8184859B2 (en) | Road marking recognition apparatus and method | |
| JP3169483B2 (en) | Road environment recognition device | |
| EP0827127B1 (en) | Local positioning apparatus, and method therefor | |
| CN102722711B (en) | License plate character sequence identification method and device in license plate type identification | |
| US7561721B2 (en) | System and method for range measurement of a preceding vehicle | |
| WO2016127545A1 (en) | Character segmentation and recognition method | |
| Daigavane et al. | Road lane detection with improved canny edges using ant colony optimization | |
| KR100975749B1 (en) | Method for recognizing lane and lane departure with Single Lane Extraction | |
| US7561720B2 (en) | Single camera system and method for range and lateral position measurement of a preceding vehicle | |
| CN105303189A (en) | Method and device for detecting specific identification image in predetermined area | |
| WO2019172059A1 (en) | Lane marker recognition device | |
| US20070031008A1 (en) | System and method for range measurement of a preceding vehicle | |
| CN101454814A (en) | Vehicle type determination device, program, and method | |
| CN105488501A (en) | Method for correcting license plate slant based on rotating projection | |
| WO2023279966A1 (en) | Multi-lane-line detection method and apparatus, and detection device | |
| CN109886175B (en) | Method for detecting lane line by combining straight line and circular arc | |
| CN104700072A (en) | Lane line historical frame recognition method | |
| CN109558765B (en) | Automobile and lane line detection method and device |
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: 20120613 |