CN101567041A - Method for recognizing characters of number plate images of motor vehicles based on trimetric projection - Google Patents
Method for recognizing characters of number plate images of motor vehicles based on trimetric projection Download PDFInfo
- Publication number
- CN101567041A CN101567041A CNA2009100270349A CN200910027034A CN101567041A CN 101567041 A CN101567041 A CN 101567041A CN A2009100270349 A CNA2009100270349 A CN A2009100270349A CN 200910027034 A CN200910027034 A CN 200910027034A CN 101567041 A CN101567041 A CN 101567041A
- Authority
- CN
- China
- Prior art keywords
- character
- binaryzation
- gray
- number plate
- value
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Character Discrimination (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to a method for recognizing deformed license plate characters of number plate images of motor vehicles. The method comprises the following steps: converting number plate color images of the motor vehicles into gray level images; performing binaryzation and character segmentation; performing trimetric projection on fuzzy binarized character images in the horizontal direction, the vertical direction and the spatial overlooking direction; forming character characteristic values through calculation; and quickly matching the character characteristic values with character characteristic values in a standard word stock so as to accurately recognize fuzzy characters. The method has the advantages that: the method can quickly match the deformed characters of the number plate images of the motor vehicles, and is used for quick and fuzzy recognition of the deformed number plate characters in the process of automatic recognition of the number plate images of the motor vehicles so as to automatically extract number of number plates of the motor vehicles for images recorded by a video image monitoring device for monitoring illegal traffic activities of the motor vehicles.
Description
Technical field
The present invention relates to a kind of recognition methods of car plate deformed characters of automotive number plate image, specifically be based on the character identifying method of tripleplane.
Background technology
According to statistics, the automobile traffic illegal activities are to cause one of high basic reason of road traffic accident.For this reason, various places vehicle supervision department of public security organ bans or prevents the automobile traffic illegal activities that a large amount of automobile traffic illegal activities video frequency graphic monitoring equipment has been installed on its area under one's jurisdiction road.Yet, quick growth along with mileages of transport route and vehicle possess amount, automobile traffic illegal activities video frequency graphic monitoring equipment is installed quantity sharply to be increased, vehicle supervision department of public security organ can not dispose timely and effectively to the image of automobile traffic illegal activities video monitoring image equipment records, now become the hot issue that broad masses complain, the automatic recognition function of configuration automotive number plate image can be accelerated the processing of automobile traffic illegal activities video image in automobile traffic illegal activities video frequency graphic monitoring equipment.But because automobile traffic illegal activities video image is subjected to the influence of factors such as mounting condition and running environment easily, often there is defective in captured video image, particularly the lack of standardization or distortion of the character on the car plate has become automatically one of technological difficulties of identification of automotive number plate image, has a strong impact on the raising of automotive number plate image recognition accuracy rate.
Summary of the invention
The present invention seeks in order to overcome the deficiencies in the prior art, accurately discern characters on license plate information in the automotive number plate image that suffers environmental pollutions such as muddy water, light, friction, a kind of character identifying method based on tripleplane is provided, in the automatic identifying of automotive number plate image,, can be used for the automatic extractor motor-car of image brand number to the equipment records of automobile traffic illegal activities video frequency graphic monitoring to the quick fuzzy diagnosis of car plate deformed characters.
According to technical scheme provided by the invention, comprise the steps: based on the character identifying method of the automotive number plate image of tripleplane
(1) 24 coloured images of automotive number plate is converted into the gray-scale map of 8 colors;
(2) described gray-scale map is carried out character cutting and binaryzation, form character cutting gray-scale map and character binaryzation figure: the character cutting gray-scale map is to divide and cut the formed character picture of number plate gray-scale map based on the picture size ratio according to each character size that takes up space in the car plate standard size; Character binaryzation figure adds up the quantity that gray-scale value 0～255 occurs in the character cutting gray-scale map earlier, calculate the threshold value of gray-scale map based on the grey level histogram theory, when gray values of pixel points surpasses or equals threshold value, this gray-scale value is classified as " 1 ", otherwise be " 0 ", each pixel is repeated aforesaid operations, thereby obtain character binaryzation figure.
(3) character binaryzation figure is carried out left and right directions, above-below direction, space and overlook direction tripleplane, the character feature value in formed character feature value and the standard character library is compared, thereby identifies the automotive number plate character information in the image.
The method of described tripleplane and comparison comprises the steps:
(1) to single character binaryzation figure according to 40 * 40 dot matrix standardization;
(2) the character binaryzation figure after the standardization is carried out projection according to the left and right horizontal direction, and the projection value of every row is inserted among the array X (40), array X (40) expression has the one-dimension array of 40 elements;
(3) the character binaryzation figure after the standardization is carried out projection according to vertical direction up and down, and the projection value of every row is inserted among the array Y (40), array Y (40) expression has the one-dimension array of 40 elements;
(4) the add up numerical value of array X (40) and Y (40) correspondence position is compared with the numerical value that is provided with in the standard character library, if coupling accurately, is then finished this character picture identification, changes step 7;
(5) the character binaryzation figure after the standardization is overlooked 4 * 4 zones that direction is divided into same size according to the space, calculate the quantity of each regional gray-scale value then respectively for " 1 ", and deposit among the array Z (16), compare with the numerical value that is provided with in the standard character library, if coupling accurately, then finish this character picture identification, change step 7; Otherwise this character picture recognition failures;
(6) statistics character binaryzation figure recognition failures number of times if the character binaryzation figure frequency of failure surpasses setting value, then finishes this image recognition;
(7) obtain corresponding characters information in the standard character library; If there is not character binaryzation figure to be identified, then finishes this image recognition, otherwise continue the character late binary picture.
Advantage of the present invention is: because deformed characters recognition methods of the present invention and stroke have nothing to do, thus provide the foundation for the correct fuzzy diagnosis of deformed characters, thus the accuracy rate of automotive number plate image Recognition of License Plate Characters improved greatly.
Description of drawings
Fig. 1 is the gray-scale map of automotive number plate image of the present invention.
Fig. 2 is the character cutting gray-scale map of automotive number plate image and the contrast figure of character binaryzation figure.
Fig. 3 is automotive number plate character binaryzation figure and corresponding characters recognition result figure.
Embodiment
The invention will be further described below in conjunction with drawings and Examples.
Be illustrated in figure 1 as the gray-scale map of automotive number plate image, the character on the image is very fuzzy, and particularly the Chinese character on the car plate can not be told Chinese character stroke, and the adhesion between the character is very serious, and this discerns for accurate character automatically and has brought difficulty.But according to recognition methods of the present invention, with ambiguous character binaryzation figure about (level) direction, (vertically) direction, space are overlooked direction and are carried out tripleplane up and down, calculate thus and form character feature value to be identified, then character feature value to be identified and standard character eigenwert are mated fast, thereby accurately identify ambiguous character.Wherein, the gray-scale map of automotive number plate image is transformed by 24 coloured images of automotive number plate.Owing to the single pixel value in 24 cromograms stores according to three kinds of colours of red, green, blue, so get a numerical value in three colours of red, green, blue, formed new number plate view data again according to the pixel order, promptly form the grey chromatic graph of 8 colors.
The top is the character cutting gray-scale map of automotive number plate image as shown in Figure 2, and the below is the character binaryzation figure of automotive number plate image.The character cutting gray-scale map is to divide and cut the formed character picture of number plate gray-scale map based on the picture size ratio according to the space size of character in the car plate standard size; Character binaryzation figure adds up 0～255 quantity that occurs in the character cutting gray-scale map earlier, calculate the threshold value of gray-scale map based on the grey level histogram theory, when gray values of pixel points surpasses or equals threshold value, this gray-scale value is classified as " 1 ", otherwise be " 0 ", each pixel is repeated aforesaid operations, thereby obtain character binaryzation figure.Wherein, histogram is the function of gray level, has the number of the pixel of every kind of gray level in the presentation video, every kind of frequency that gray scale occurs in the reflection image; Threshold value is specific gray-scale value.
The top is character binaryzation figure as shown in Figure 3, and the below is character identification result figure.(level) direction about carrying out according to the character binaryzation figure to the automotive number plate image, direction tripleplane is overlooked in (vertically) direction, space up and down, character feature value in formed character feature value and the standard character library is compared fast, thereby identify the automotive number plate character information, the concrete operations step is:
1, to the character binaryzation figure that obtains according to 40 * 40 dot matrix standardization, be character binaryzation figure unification not of uniform size the standard character image of 40 * 40 dot matrix promptly so that with character library in character feature compare.
2, with the character binaryzation figure after the standardization according to about (level) direction carry out projection, the gray-scale value that calculates every row is the pixel quantity of " 1 ", and inserts among the array X (40).
3, with the character binaryzation figure after the standardization according to about (vertically) direction carry out projection, the gray-scale value that calculates every row is the pixel quantity of " 1 ", and inserts among the array Y (40).
4, the numerical value addition of array X (40) and Y (40) correspondence position, compare fast with the numerical value that is provided with in the standard character library.If coupling accurately, is then finished this character picture identification, change step 7.Described coupling accurately means matching rate and reaches more than 98%, and concrete matching rate can be set according to recognition effect, down together.
5, the character binaryzation figure after the standardization is overlooked 4 * 4 zones that direction is divided into same size according to the space, calculate 16 zones then respectively and the quantity of gray-scale value occurs for " 1 ", and deposit among the array Z (16), compare fast with the numerical value that is provided with in the standard character library.If coupling accurately, is then finished this character picture identification, change step (7); Otherwise this character picture recognition failures.
Described array X (40) and Y (40) all represent to have the one-dimension array of 40 elements.Array Z (16) expression has the one-dimension array of 16 elements.
For step 5,4 * 4 zones are unidentified to go out character if divide, and can also divide 8 * 8 zones and discern.
6, the statistics character binaryzation figure quantity of failing if the character binaryzation figure frequency of failure surpasses 3 times, then finishes this number plate image recognition.
7, obtain corresponding characters information in the standard character library,, then finish this number plate image recognition, otherwise continue the identification of character late binary picture if there is not character binaryzation figure to be identified.
The invention solves the fuzzy quick identification difficult problem after the characters on license plate distortion on the automotive number plate image, for accurate and effective extractor motor-car log-on message provides the foundation, for the automatic recognition function of widespread adoption automotive number plate on automobile traffic illegal activities video frequency graphic monitoring equipment, the processing of accelerating automobile traffic illegal activities video image provide technical support.
Claims (2)
1, based on the character identifying method of the automotive number plate image of tripleplane, it is characterized in that: described method comprises the steps:
(1) 24 coloured images of automotive number plate is converted into the gray-scale map of 8 colors;
(2) described gray-scale map is carried out character cutting and binaryzation, form character cutting gray-scale map and character binaryzation figure: the character cutting gray-scale map is to divide and cut the formed character picture of number plate gray-scale map based on the picture size ratio according to each character size that takes up space in the car plate standard size; Character binaryzation figure adds up the quantity that gray-scale value 0～255 occurs in the character cutting gray-scale map earlier, calculate the threshold value of gray-scale map based on the grey level histogram theory, when gray values of pixel points surpasses or equals threshold value, this gray-scale value is classified as " 1 ", otherwise be " 0 ", each pixel is repeated aforesaid operations, thereby obtain character binaryzation figure.
(3) character binaryzation figure is carried out left and right directions, above-below direction, space and overlook direction tripleplane, the character feature value in formed character feature value and the standard character library is compared, thereby identifies the automotive number plate character information in the image.
2, the character identifying method of the automotive number plate image based on tripleplane as claimed in claim 1 is characterized in that the method for described tripleplane and comparison comprises the steps:
(1) to single character binaryzation figure according to 40 * 40 dot matrix standardization;
(2) the character binaryzation figure after the standardization is carried out projection according to the left and right horizontal direction, and the projection value of every row is inserted among the array X (40), array X (40) expression has the one-dimension array of 40 elements;
(3) the character binaryzation figure after the standardization is carried out projection according to vertical direction up and down, and the projection value of every row is inserted among the array Y (40), array Y (40) expression has the one-dimension array of 40 elements;
(4) the add up numerical value of array X (40) and Y (40) correspondence position is compared with the numerical value that is provided with in the standard character library, if coupling accurately, is then finished this character picture identification, changes step 7;
(5) the character binaryzation figure after the standardization is overlooked 4 * 4 zones that direction is divided into same size according to the space, calculate the quantity of each regional gray-scale value then respectively for " 1 ", and deposit among the array Z (16), compare with the numerical value that is provided with in the standard character library, if coupling accurately, then finish this character picture identification, change step 7; Otherwise this character picture recognition failures;
(6) statistics character binaryzation figure recognition failures number of times if the character binaryzation figure frequency of failure surpasses setting value, then finishes this image recognition;
(7) obtain corresponding characters information in the standard character library; If there is not character binaryzation figure to be identified, then finishes this image recognition, otherwise continue the character late binary picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2009100270349A CN101567041A (en) | 2009-05-25 | 2009-05-25 | Method for recognizing characters of number plate images of motor vehicles based on trimetric projection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNA2009100270349A CN101567041A (en) | 2009-05-25 | 2009-05-25 | Method for recognizing characters of number plate images of motor vehicles based on trimetric projection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101567041A true CN101567041A (en) | 2009-10-28 |
Family
ID=41283188
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA2009100270349A Pending CN101567041A (en) | 2009-05-25 | 2009-05-25 | Method for recognizing characters of number plate images of motor vehicles based on trimetric projection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101567041A (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101901479A (en) * | 2010-07-26 | 2010-12-01 | 公安部交通管理科学研究所 | Method for optimizing license plate image with character color dropout |
| CN101901478A (en) * | 2010-07-26 | 2010-12-01 | 公安部交通管理科学研究所 | Image definition enhancing method for treating color fading of license plate |
| CN103116984A (en) * | 2013-01-21 | 2013-05-22 | 信帧电子技术（北京）有限公司 | Method to detect illegal parking |
| CN106599894A (en) * | 2016-12-27 | 2017-04-26 | 上海铁路局科学技术研究所 | Method for identifying pole number of overhead line system based on image identification |
| CN107273890A (en) * | 2017-05-26 | 2017-10-20 | 亿海蓝（北京）数据技术股份公司 | Graphical verification code recognition methods and device for character combination |
| CN107452144A (en) * | 2017-08-17 | 2017-12-08 | 成都工业学院 | Automatic charging method and device |
| WO2018086233A1 (en) * | 2016-11-08 | 2018-05-17 | 广州视源电子科技股份有限公司 | Character segmentation method and device, and element detection method and device |
| CN109598271A (en) * | 2018-12-10 | 2019-04-09 | 北京奇艺世纪科技有限公司 | A kind of character segmentation method and device |
| CN110020442A (en) * | 2019-04-12 | 2019-07-16 | 上海电机学院 | A kind of portable translating machine |
| CN110659632A (en) * | 2019-09-29 | 2020-01-07 | 公安部交通管理科学研究所 | System and method for testing motor vehicle number plate identification performance of traffic technology monitoring equipment based on image block assignment |
| CN110781901A (en) * | 2019-10-29 | 2020-02-11 | 湖北工业大学 | Instrument ghost character recognition method based on BP neural network prediction threshold |
| CN111241344A (en) * | 2020-01-14 | 2020-06-05 | 新华智云科技有限公司 | Video duplicate checking method, system, server and storage medium |
-
2009
- 2009-05-25 CN CNA2009100270349A patent/CN101567041A/en active Pending
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101901479A (en) * | 2010-07-26 | 2010-12-01 | 公安部交通管理科学研究所 | Method for optimizing license plate image with character color dropout |
| CN101901478A (en) * | 2010-07-26 | 2010-12-01 | 公安部交通管理科学研究所 | Image definition enhancing method for treating color fading of license plate |
| CN103116984A (en) * | 2013-01-21 | 2013-05-22 | 信帧电子技术（北京）有限公司 | Method to detect illegal parking |
| CN103116984B (en) * | 2013-01-21 | 2016-03-23 | 信帧电子技术（北京）有限公司 | Detect the method for parking offense |
| WO2018086233A1 (en) * | 2016-11-08 | 2018-05-17 | 广州视源电子科技股份有限公司 | Character segmentation method and device, and element detection method and device |
| CN106599894A (en) * | 2016-12-27 | 2017-04-26 | 上海铁路局科学技术研究所 | Method for identifying pole number of overhead line system based on image identification |
| CN107273890A (en) * | 2017-05-26 | 2017-10-20 | 亿海蓝（北京）数据技术股份公司 | Graphical verification code recognition methods and device for character combination |
| CN107452144A (en) * | 2017-08-17 | 2017-12-08 | 成都工业学院 | Automatic charging method and device |
| CN109598271A (en) * | 2018-12-10 | 2019-04-09 | 北京奇艺世纪科技有限公司 | A kind of character segmentation method and device |
| CN110020442A (en) * | 2019-04-12 | 2019-07-16 | 上海电机学院 | A kind of portable translating machine |
| CN110659632A (en) * | 2019-09-29 | 2020-01-07 | 公安部交通管理科学研究所 | System and method for testing motor vehicle number plate identification performance of traffic technology monitoring equipment based on image block assignment |
| CN110659632B (en) * | 2019-09-29 | 2022-08-12 | 公安部交通管理科学研究所 | System and method for testing motor vehicle number plate identification performance of traffic technology monitoring equipment based on image block assignment |
| CN110781901A (en) * | 2019-10-29 | 2020-02-11 | 湖北工业大学 | Instrument ghost character recognition method based on BP neural network prediction threshold |
| CN110781901B (en) * | 2019-10-29 | 2023-04-28 | 湖北工业大学 | Instrument ghost character recognition method based on BP neural network prediction threshold |
| CN111241344A (en) * | 2020-01-14 | 2020-06-05 | 新华智云科技有限公司 | Video duplicate checking method, system, server and storage medium |
| CN111241344B (en) * | 2020-01-14 | 2023-09-05 | 新华智云科技有限公司 | Video duplicate checking method, system, server and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101567041A (en) | Method for recognizing characters of number plate images of motor vehicles based on trimetric projection | |
| CN106373426B (en) | Parking stall based on computer vision and violation road occupation for parking monitoring method | |
| CN106991820B (en) | Illegal vehicle processing method and device | |
| CN110298300B (en) | Method for detecting vehicle illegal line pressing | |
| CN111161543A (en) | Automatic snapshot method and system for bus front violation behavior based on image recognition | |
| CN103824037B (en) | Vehicle anti-tracking alarm device | |
| CN106886987A (en) | A kind of train license plate binary image interfusion method | |
| CN102880863B (en) | Method for positioning license number and face of driver on basis of deformable part model | |
| CN107016362B (en) | Vehicle weight recognition method and system based on vehicle front windshield pasted mark | |
| CN101567042B (en) | Method for recognizing characters of number plate images of armed police automobiles | |
| CN110619279A (en) | Road traffic sign instance segmentation method based on tracking | |
| CN106919910B (en) | Traffic sign identification method based on HOG-CTH combined features | |
| CN102880859A (en) | Method for recognizing number plate | |
| CN105354533B (en) | A kind of unlicensed vehicle model recognizing method of bayonet based on bag of words | |
| CN111640293A (en) | Deep learning non-motor vehicle lane driving detection method based on embedded terminal | |
| KR102240266B1 (en) | Apparatus, system and method for classifying vehicles using image analysis | |
| CN103927548A (en) | Novel vehicle collision avoiding brake behavior detection method | |
| CN112115939A (en) | Vehicle license plate recognition method and device | |
| CN101931792A (en) | Multi-plate number locating method in high definition video steaming | |
| CN102169583B (en) | Vehicle shielding detection and segmentation method based on vehicle window positioning | |
| CN105551062A (en) | Night object detection method | |
| Chowdhury et al. | An adaptive technique for computer vision based vehicles license plate detection system | |
| CN103235940A (en) | Method for detecting multiple license plates in high-resolution image sequence | |
| CN113487878A (en) | Motor vehicle illegal line pressing running detection method and system | |
| CN110969135B (en) | Vehicle logo recognition method in natural scene |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Open date: 20091028 |