CN105117726A - License plate positioning method based on multi-feature area accumulation - Google Patents
License plate positioning method based on multi-feature area accumulation Download PDFInfo
- Publication number
- CN105117726A CN105117726A CN201510481922.3A CN201510481922A CN105117726A CN 105117726 A CN105117726 A CN 105117726A CN 201510481922 A CN201510481922 A CN 201510481922A CN 105117726 A CN105117726 A CN 105117726A
- Authority
- CN
- China
- Prior art keywords
- region
- image
- license plate
- car plate
- area
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000009825 accumulation Methods 0.000 title claims abstract description 26
- 238000000605 extraction Methods 0.000 claims abstract description 8
- 238000012937 correction Methods 0.000 claims abstract description 3
- 238000001914 filtration Methods 0.000 claims description 17
- 238000001514 detection method Methods 0.000 claims description 13
- 239000000284 extract Substances 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 7
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000003709 image segmentation Methods 0.000 claims description 3
- 238000013316 zoning Methods 0.000 claims description 3
- 230000001186 cumulative effect Effects 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 12
- 238000012216 screening Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000007726 management method Methods 0.000 description 4
- 230000000877 morphologic effect Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000005286 illumination Methods 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 230000004807 localization Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/24—Aligning, centring, orientation detection or correction of the image
- G06V10/245—Aligning, centring, orientation detection or correction of the image by locating a pattern; Special marks for positioning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/62—Text, e.g. of license plates, overlay texts or captions on TV images
- G06V20/625—License plates
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Character Input (AREA)
- Image Processing (AREA)
Abstract
The invention provides a license plate positioning method based on multi-feature area accumulation. The method comprises the following steps of pretreatment; area extraction; license-plate coarse positioning; rotation correction and accurate positioning. In the invention, an advantage of a current mainstream algorithm is combined and respective corresponding disadvantages are overcome; an area accumulation license plate screening strategy is provided so that the method can be adapted to complex environments of a stained license plate, a night environment, multiple scenes and the like. An actual application demand can be satisfied.
Description
Technical field
The present invention relates to a kind of license plate locating method, be specifically related to a kind of license plate locating method based on multiple features zone-accumulation.
Background technology
License plate recognition technology is an important application of Current vehicle intellectual management, and along with the number of vehicle constantly increases sharply, bring great inconvenience to the management of vehicle, in addition the increase of cost of labor, traditional labor management does not meet present demand.License plate recognition technology be utilize camera collection to the image comprising license board information, utilize the technology such as image procossing, pattern-recognition license board information therefrom to be extracted.Car license recognition is as the core of vehicle management system, and its discrimination is the focus that numerous researchers pay close attention to always, the Car license recognition particularly under different scene.Because original image is generally collected in outdoor physical environment, generally can be subject to the factor such as stained, deformation of illumination, the impact of sleety weather and car plate itself, the difficulty of Car license recognition is increased greatly.In general, Car license recognition is mainly divided into three parts: License Plate, Character segmentation and character recognition.Wherein License Plate is the basis of whole Car license recognition is also core, and the correctness of location decides the discrimination of whole system.Certainly, the follow-up step of License Plate failure is lost meaning.Therefore, License Plate attracts numerous scholar as the core of license plate recognition technology always and endeavours research, the algorithm of numerous License Plate that is born.Sum up, mainly contain following a few class: based on morphologic License Plate, it utilizes the character texture characteristics of car plate itself, mathematical morphology operation is used (to expand, corrode, opening/closing computing etc.) be easy to the region finding car plate place, coordinate certain candidate regions screening conditions, the particular location of car plate can be found.But the structural element that its deficiency is its mathematical morphology itself can not self-adaptation, not strong for the License Plate robustness (Robust) under different scene; Easily cause license plate area and other noise edge adhesions based on morphologic method, increase location probability of failure; When texture information is more weak, at particularly night, often locate failure based on morphologic localization method.Based on the method for the License Plate of sweep trace, its principle detects Gray Level Jump information to pretreated bianry image sweep trace at certain intervals (level, vertical direction).Record the alternatively district, region of frequent saltus step, finally determine the position of final car plate in conjunction with car plate priori.The method has and realizes simple, fireballing feature, but is only applicable to the Car license recognition under the desirable or simple scenario of binaryzation, to complex scene or the stained situation adaptability of car plate not strong.License Plate based on color is the priori utilizing the background of China's car plate and the constant color combination of character, carries out specific color lookup, finally obtains color of object block, thus determine car plate position to image.In the ordinary course of things (uniform illumination, car plate are cleaner), the method accurate positioning is used.The scene of inapplicable dark-coloured car plate or night colors distortion.
Summary of the invention
The object of this invention is to provide the license plate locating method based on multiple features zone-accumulation, make it possible to adapt to the complex environments such as stained, the night-environment of car plate and many scenes.The demand of practical application can be competent at.
The invention provides following technical scheme:
Based on the license plate locating method of multiple features zone-accumulation, comprise the steps:
Pre-service;
Extracted region;
Car plate coarse positioning;
Rotational correction;
Accurate location.
Further, pre-service comprises: gray processing, binaryzation and rim detection;
Gray processing, the formula turning gray-scale map based on RGB color space is as follows:
Gray(i，j)＝R(i，j)*0.299+G(i，j)*0.587+B(i，j)*0.114(1)
Wherein Gray (i, j) is the gray-scale value of target image on point (i, j), and R, G, B are image three components;
Binaryzation: gray-scale map adopts the gray level of 0-255 to represent piece image, and pixel gradient is also not obvious, is therefore unfavorable for the extraction to Image Segmentation Using and edge; Meanwhile, in order to reduce unnecessary information and calculated amount, we only use 0 and 255 two value to represent the target that we pay close attention to; Generally, 0 represents background, the target that 255 representatives are paid close attention to; The computing formula of binaryzation is as follows:
Rim detection: adopt Sobel operator can extract the marginal information of horizontal direction or vertical direction as required.
Further, extracted region comprises: the connected domain based on edge feature is extracted, extract based on the connected domain of background color feature, extract and forming region cumulative chart based on the connected domain of character colour characteristic.
Further, the connected domain extraction based on edge feature comprises the steps:
Step1. Top-Hat conversion is performed to gray-scale map, remove highlight area;
Step2. use Sobel operator to detect vertical edge, destroy the line segment of horizontal direction;
Step3. edge figure carries out denoising; Specific practice is that the profile retrieving outline map all is designated as C, and any Ci asks for its area A rea (Ci); If Area (Ci) <minAreaThresh or Area (Ci) >maxAreaThresh, use black picture element to fill this profile, otherwise retain; Wherein minAreaThreash and maxAreaThresh represents that the car plate of permission is minimum, maximum area.
Step4. construct suitable structural element, perform closed operation, opening operation operates; The size of structural element is generally set as 22 × 4, and occurrence also can experimentally result be finely tuned;
Step5. the connected domain of above-mentioned steps is screened; Find out all connected domains of image, obviously do not meet the region of vehicle license plate characteristic according to condition filterings such as car plate area, wide high proportions;
Step6. by above-mentioned steps, candidate regions S1 is finally obtained.
Further, the connected domain extraction based on background color feature comprises the steps:
Step1. by original image from RGB Color Color spatial alternation to hsv color space, RGB changes the computing formula of HSV and is:
v＝max；
Step2. detect the intrinsic background color of car plate, traversing graph is as each pixel, if this pixel is the inherent colour of car plate, then setting this value is 255, otherwise is 0;
Step3. above-mentioned image is expanded, make white pixel be linked to be region;
Step4. the above area of filtering, the ungratified connected domain of wide high proportion;
Step5. by above-mentioned steps, candidate regions S2 is finally obtained.
Further, the connected domain based on character colour characteristic is extracted, and concrete steps are as follows:
Step1. detect car plate character color to the hsv color figure of image, travel through each pixel of this image, if object pixel, then setting this value is 255, otherwise is 0;
Step2. to upper figure binaryzation, and expansion plans picture, make character color pixel be linked to be region;
Step3. filtering area, the ungratified region of the ratio of width to height;
Step4. by above-mentioned steps, candidate regions S3 is finally obtained.
Further, zone-accumulation concrete grammar is as follows:
Step1. single channel figure f (x, y) equal with original image size M × N is created, and by its all pixel value zero-setting operation, namely
Step2. the candidate regions S1 will extracted, S2, S3 are included into Candidate Set S, by each region S
ibe made the single channel mask Mask that a pixel value is p
i(x, y);
Step3. to image f (x, y),
wherein x0, y0, w, h are respectively region Si in the origin coordinates of image space and wide high level;
Step4. to above-mentioned image f (x, y), the layer of zoning superposition
If Step5. layer=0, then represent in all candidate regions, there is no overlapping region, now only get the region the most similar to car plate in S1 as final license plate area; If layer=1, all in filtering S do not have overlapping region, i.e. Overlap (S)=0; If layer > 1, the in like manner region of filtering Overlap (S)≤1;
Step6. binaryzation is carried out to image f (x, y), wherein threshold value
finally select the region of car plate.
The invention has the beneficial effects as follows: the multiple features such as the background that methods combining car plate texture, background color and car plate are intrinsic and character color collocation position car plate.Morphology is utilized to get the candidate region of car plate; The method of searching color is utilized to find out the connected domain block of car plate candidate block and character color fusion, then use the comparatively obvious non-license plate area that a kind of method filtering of zone-accumulation is all, then find car plate region roughly in conjunction with the feature of car plate own.Finally, an optimum complete license plate area is obtained according to the region superposition in this approximate region.Based on the methods combining advantage of current main flow algorithm of the polymorphic type License Plate of multiple features zone-accumulation, overcome corresponding shortcoming separately, and a kind of zone-accumulation car plate screening strategy is proposed, make it possible to adapt to the complex environments such as stained, the night-environment of car plate and many scenes.The demand of practical application can be competent at.
Accompanying drawing explanation
Accompanying drawing is used to provide a further understanding of the present invention, and forms a part for instructions, together with embodiments of the present invention for explaining the present invention, is not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is structural representation of the present invention;
Fig. 2 is the structural representation of a preferred embodiment;
Fig. 3 is Sobel operator edge extracting figure;
Fig. 4 extracts car plate process flow diagram based on zone-accumulation;
Fig. 5 is the candidate region based on textural characteristics;
Fig. 6 is the candidate region based on color characteristic;
Fig. 7 is zone-accumulation figure.
Embodiment
As shown in figs. 1-7, being positioned such that in the present embodiment for car plate:
Pre-service, the image arrived due to camera collection, often based on the color diagram of RGB, in order to reduce the computation complexity of image recognition, often needs the color notation conversion space of image to gray space.Pretreated object is mainly removed or is reduced the noise of non-license plate area, retains or strengthen the structure feature information of license plate area.Pre-service mainly comprises the steps such as gray processing, binaryzation, rim detection.
Gray processing: camera is when collection image, and the data gathered are generally the coloured images of the 3-passage (channels) based on RGB color space.Although coloured image can preserve more information, must to sacrifice internal memory and arithmetic speed for cost.In the location of car plate, needing the operation carried out not need often to process under color diagram, in order to meet the requirement of real-time, needing image to carry out gray processing process.
Binaryzation: gray-scale map adopts the gray level of 0-255 to represent piece image, and pixel gradient is also not obvious, is therefore unfavorable for the extraction to Image Segmentation Using and edge.Meanwhile, in order to reduce unnecessary information and calculated amount, we only use 0 and 255 two value to represent the target that we pay close attention to.Generally, 0 represents background, the target that 255 representatives are paid close attention to.The computing formula of binaryzation is as follows:
The main difficult point of image binaryzation is the selection of threshold value.Under general situation, have the selection of two kinds of threshold values, one is the fixed threshold based on the overall situation, and two is Adaptive Thresholding.The former computing method are the mean value obtaining gray-scale map, then fluctuate on the basis of mean value, and the concrete numerical value that floats is determined with reference to the environment of reality and the effect of binaryzation.Adaptive algorithm adopts Da-Jin algorithm (OTSU) to ask threshold value.OTSU algorithm main thought is that the grey of image is divided into 2 parts by gray level, and make the grey value difference between two parts maximum, the gray difference between each part is minimum, finds a suitable grey level divide by the calculating of variance.The simple and reliable property of OTSU algorithm calculated threshold is also higher.
Rim detection, the algorithm of rim detection conventional in Iamge Segmentation has Canny operator, Sobel operator, Laplace operator and Robert operator etc.Relative to other operators, Sobel operator can extract the marginal information of horizontal direction or vertical direction as required, and the skirt response effect of edge in the horizontal direction and vertical direction is best.According to the textural characteristics of car plate itself, the marginal information of its vertical direction is rule distribution, is beneficial to the extraction of license plate area.Therefore the present invention adopts Sobel operator as rim detection.Perform rim detection effect as shown in Figure 3.
In a general case, the texture information of car plate is as a key character of License Plate, by extracting the marginal information of car plate, utilize mathematical morphology to carry out a series of expansion, corrosion, the operation such as opening/closing computing can get the multiple candidate's connected domain of car plate.This method is simple also efficient, but some shortcomings existed due to morphological operation (as structural element can not self-adaptation, easily with other noise adhesions etc.) make License Plate in some scenarios failed often.Therefore, texture information is only utilized obviously can not to reach requirement for the License Plate of different scene in practical application.The present invention, based on beyond textural characteristics, increases the consideration to car plate background and characters on license plate color characteristic.Extract the region of car plate background color and the region of characters on license plate color, a kind of method based on zone-accumulation then using the present invention to propose filters out car plate.Method And Principle based on zone-accumulation is the accumulation superposition carried out in the region that the feature based on texture, color and character three parts is extracted spatially, according to the number of plies superposed all non-license plate area filterings, finally navigates to the physical location of car plate.
Connected domain based on edge feature is extracted:
The Edge texture information of car plate is a key character of car plate.License plate area texture-rich and have certain regularity of distribution, what utilize that this feature can be very fast determines car plate position.But in the application of reality, the scene that Car license recognition uses is often very complicated, also just means the how effective filter out background noise of needs consideration, thus the candidate regions of class car plate is extracted.The way taked in the present embodiment is as follows:
Step1. Top-Hat conversion is performed to gray-scale map, remove highlight area.
Step2. use Sobel operator to detect vertical edge, destroy the line segment of horizontal direction.
Step3. edge figure carries out denoising.Specific practice is that the profile retrieving outline map all is designated as C, and any Ci asks for its area A rea (Ci).If Area (Ci) <minAreaThresh or Area (Ci) >maxAreaThresh, use black picture element to fill this profile, otherwise retain.Wherein minAreaThreash and maxAreaThresh represents that the car plate of permission is minimum, maximum area.
Step4. construct suitable structural element, perform closed operation, opening operation operates.The size of structural element is generally set as 22 × 4, and occurrence also can experimentally result be finely tuned.
Step5. the connected domain of above-mentioned steps is screened.Find out all connected domains of image, obviously do not meet the region of vehicle license plate characteristic according to condition filterings such as car plate area, wide high proportions.
Step6. by above-mentioned steps, candidate regions S1 is finally obtained.
Connected domain based on background color feature is extracted:
Because the background color of China's car plate is relatively fixing, the license plate locating method therefore based on color characteristic also occupies very large status in location algorithm.The present invention determines the regional location of car plate by the background colour that car plate in detection color diagram is intrinsic.Specific algorithm step is as follows:
Step1. by original image from RGB Color Color spatial alternation to hsv color space.The computing formula that RGB changes HSV is:
v＝max
Step2. the intrinsic background color (being assumed to be blueness) of car plate is detected.Traversing graph is as each pixel, if this pixel is blue, then setting this value is 255, otherwise is 0
Step3. above-mentioned image is expanded, make white pixel be linked to be region.
Step4. the above area of filtering, the ungratified connected domain of wide high proportion.
Step5. by above-mentioned steps, candidate regions S2 is finally obtained.
Connected domain based on character colour characteristic is extracted:
In some cases, as car plate paler colour or color distortion, only be sometimes difficult to the region of car plate to be found to the detection of car plate background color.In order to make Car license recognition have higher applicability, need to take in these scenes.Except the detection of above-mentioned car plate background, the detection of character color has features such as being evenly distributed, texture information is stable equally.The present invention determines license plate area by detecting the distinctive color of characters on license plate, and specific algorithm step is as follows:
Step1. car plate character color (being assumed to be white) is detected to the hsv color figure of image.Travel through each pixel of this image, if object pixel, then setting this value is 255, otherwise is 0.
Step2. to upper figure binaryzation, and expansion plans picture, make white pixel be linked to be region.
Step3. filtering area, the ungratified region of the ratio of width to height.
Step4. by above-mentioned steps, candidate regions S3 is finally obtained.
Zone-accumulation figure:
Car plate under different scenes, under the condition such as stained in illumination, car plate, inclination.The situation that car plate is imperfect, license plate area ruptures may be there is in the license plate area obtained after its pre-service.The present invention is by the license plate candidate area S1 of above three steps, and the extraction of S2, S3, real license plate area therefrom extracts by principle accurately that utilize region to superpose, and the car plate of extracted region optimum in the superposition of car plate.Zone-accumulation specific algorithm is as follows:
Step1. single channel figure f (x, y) equal with original image size M × N is created, and by its all pixel value zero-setting operation, namely
Step2. the candidate regions S1 will extracted, S2, S3 are included into Candidate Set S, by each region S
ibe made the single channel mask Mask that a pixel value is p
i(x, y).
Step3. to image f (x, y),
wherein x0, y0, w, h are respectively region Si in the origin coordinates of image space and wide high level.
Step4. to above-mentioned image f (x, y), the layer of zoning superposition
If Step5. layer=0, then represent in all candidate regions, there is no overlapping region, now only get the region the most similar to car plate in S1 as final license plate area; If layer=1, all in filtering S do not have overlapping region, i.e. Overlap (S)=0; If layer > 1, the in like manner region of filtering Overlap (S)≤1.
Step6. binaryzation is carried out to image f (x, y).Wherein threshold value
finally select the region of car plate.
Optimum car plate selects to need to select from car plate overlapping region an optimum region, if license plate area in Fig. 7 is the brightest region of gray scale.This region is superposed by two layer regions, in order to ensure the license plate area defect caused due to a variety of causes, therefore needs to determine in the region of overlap that an optimum region is as license plate area.The strategy that the present invention takes is:
Region (LP)=(top, bottom, left, right), wherein:
The coboundary that top=gray scale is the brightest
The lower boundary that bottom=gray scale is the brightest
The minimum left margin of letf=
The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, although with reference to previous embodiment to invention has been detailed description, for a person skilled in the art, it still can be modified to the technical scheme described in foregoing embodiments, or carries out equivalent replacement to wherein portion of techniques feature.Within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (7)
1. based on a license plate locating method for multiple features zone-accumulation, it is characterized in that, comprise the steps:
Pre-service;
Extracted region;
Car plate coarse positioning;
Rotational correction;
Accurate location.
2. the license plate locating method based on multiple features zone-accumulation according to claim 1, is characterized in that, described pre-service comprises: gray processing, binaryzation and rim detection;
Gray processing, the formula turning gray-scale map based on RGB color space is as follows:
(1)
Wherein Gray (i, j) is the gray-scale value of target image on point (i, j), and R, G, B are image three components;
Binaryzation: gray-scale map adopts the gray level of 0-255 to represent piece image, and pixel gradient is also not obvious, is therefore unfavorable for the extraction to Image Segmentation Using and edge; Meanwhile, in order to reduce unnecessary information and calculated amount, we only use 0 and 255 two value to represent the target that we pay close attention to; Generally, 0 represents background, the target that 255 representatives are paid close attention to; The computing formula of binaryzation is as follows:
Rim detection: adopt Sobel operator can extract the marginal information of horizontal direction or vertical direction as required.
3. the license plate locating method based on multiple features zone-accumulation according to claim 1, it is characterized in that, extracted region comprises: the connected domain based on edge feature is extracted, extract based on the connected domain of background color feature, extract and forming region cumulative chart based on the connected domain of character colour characteristic.
4. the license plate locating method based on multiple features zone-accumulation according to claim 3, is characterized in that, the connected domain based on edge feature is extracted and comprised the steps:
Step1. Top-Hat conversion is performed to gray-scale map, remove highlight area;
Step2. use Sobel operator to detect vertical edge, destroy the line segment of horizontal direction;
Step3. edge figure carries out denoising; Specific practice is that the profile retrieving outline map all is designated as C, and any Ci asks for its area A rea (Ci); If Area (Ci) <minAreaThresh or Area (Ci) >maxAreaThresh, use black picture element to fill this profile, otherwise retain; Wherein minAreaThreash and maxAreaThresh represents that the car plate of permission is minimum, maximum area.
Step4. construct suitable structural element, perform closed operation, opening operation operates; The size of structural element is generally set as 22 × 4, and occurrence also can experimentally result be finely tuned;
Step5. the connected domain of above-mentioned steps is screened; Find out all connected domains of image, obviously do not meet the region of vehicle license plate characteristic according to condition filterings such as car plate area, wide high proportions;
Step6. by above-mentioned steps, candidate regions S1 is finally obtained.
5. the license plate locating method based on multiple features zone-accumulation according to claim 3, is characterized in that, the connected domain based on background color feature is extracted and comprised the steps:
Step1. by original image from RGB Color Color spatial alternation to hsv color space, RGB changes the computing formula of HSV and is:
Step2. detect the intrinsic background color of car plate, traversing graph is as each pixel, if this pixel is the inherent colour of car plate, then setting this value is 255, otherwise is 0;
Step3. above-mentioned image is expanded, make white pixel be linked to be region;
Step4. the above area of filtering, the ungratified connected domain of wide high proportion;
Step5. by above-mentioned steps, candidate regions S2 is finally obtained.
6. the license plate locating method based on multiple features zone-accumulation according to claim 3, is characterized in that, the connected domain based on character colour characteristic is extracted, and concrete steps are as follows:
Step1. detect car plate character color to the hsv color figure of image, travel through each pixel of this image, if object pixel, then setting this value is 255, otherwise is 0;
Step2. to upper figure binaryzation, and expansion plans picture, make character color pixel be linked to be region;
Step3. filtering area, the ungratified region of the ratio of width to height;
Step4. by above-mentioned steps, candidate regions S3 is finally obtained.
7. the license plate locating method based on multiple features zone-accumulation according to any one in claim 3 to 6, is characterized in that, zone-accumulation concrete grammar is as follows:
Step1. single channel figure f (x, y) equal with original image size M × N is created, and by its all pixel value zero-setting operation, namely
Step2. the candidate regions S1 will extracted, S2, S3 are included into Candidate Set S, each region Si are made the single channel mask Mask that a pixel value is p
i(x, y);
Step3. to image f(x, y),
Wherein x0, y0, w, h are respectively region Si in the origin coordinates of image space and wide high level;
Step4. to above-mentioned image f(x, y), the layer of zoning superposition
If Step5. layer=0, then represent in all candidate regions, there is no overlapping region, now only get the region the most similar to car plate in si as final license plate area; If layer=1, all regions not having overlap in filtering S, i.e. overlap(S)=0; If layer > 1, in like manner filtering Overlap((S))≤region of 1;
Step6. to image f(x, y) carry out binaryzation, wherein threshold value
finally select the region of car plate.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510481922.3A CN105117726B (en) | 2015-08-07 | 2015-08-07 | License plate locating method based on multiple features zone-accumulation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510481922.3A CN105117726B (en) | 2015-08-07 | 2015-08-07 | License plate locating method based on multiple features zone-accumulation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105117726A true CN105117726A (en) | 2015-12-02 |
| CN105117726B CN105117726B (en) | 2018-10-09 |
Family
ID=54665709
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510481922.3A Expired - Fee Related CN105117726B (en) | 2015-08-07 | 2015-08-07 | License plate locating method based on multiple features zone-accumulation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105117726B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778741A (en) * | 2016-12-08 | 2017-05-31 | 中国人民解放军国防科学技术大学 | A kind of license plate locating method in the complex background image shot based on handheld terminal |
| CN106815580A (en) * | 2016-12-23 | 2017-06-09 | 上海集成电路研发中心有限公司 | A kind of license plate locating method |
| CN107292898A (en) * | 2017-05-04 | 2017-10-24 | 浙江工业大学 | A kind of car plate shadow Detection and minimizing technology based on HSV |
| CN107578044A (en) * | 2017-05-08 | 2018-01-12 | 浙江工业大学 | A kind of license plate locating method based on car plate Edge texture feature |
| CN107578379A (en) * | 2017-07-12 | 2018-01-12 | 河海大学 | A kind of processing method of chess robot to checkerboard image |
| CN107609555A (en) * | 2017-09-15 | 2018-01-19 | 北京文安智能技术股份有限公司 | Detection method of license plate, model recognizing method and relevant apparatus using it |
| CN108197521A (en) * | 2017-12-31 | 2018-06-22 | 南京理工大学 | A kind of leggy Quick Response Code obtains identification device and method |
| CN108596172A (en) * | 2018-04-11 | 2018-09-28 | 航天科技控股集团股份有限公司 | A kind of the vehicle binding device and method of car networking client |
| CN110398182A (en) * | 2019-07-04 | 2019-11-01 | 西人马帝言(北京)科技有限公司 | Automatic scoring round target system, method, equipment and medium |
| CN110569833A (en) * | 2019-09-06 | 2019-12-13 | 上海应用技术大学 | License plate positioning method |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101334837A (en) * | 2008-07-31 | 2008-12-31 | 重庆大学 | Multi-method integrated license plate image positioning method |
| CN103984949A (en) * | 2014-06-11 | 2014-08-13 | 四川九洲电器集团有限责任公司 | License plate positioning method and system based on high and low cap transformation and connected domain |
-
2015
- 2015-08-07 CN CN201510481922.3A patent/CN105117726B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101334837A (en) * | 2008-07-31 | 2008-12-31 | 重庆大学 | Multi-method integrated license plate image positioning method |
| CN103984949A (en) * | 2014-06-11 | 2014-08-13 | 四川九洲电器集团有限责任公司 | License plate positioning method and system based on high and low cap transformation and connected domain |
Non-Patent Citations (2)
| Title |
|---|
| 欧阳文卫: "车牌定位算法研究", 《湖南工业职业技术学院学报》 * |
| 毛江平: "车载云台摄像机的车牌识别系统研究", 《中国优秀硕士学位论文全文数据库》 * |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106778741A (en) * | 2016-12-08 | 2017-05-31 | 中国人民解放军国防科学技术大学 | A kind of license plate locating method in the complex background image shot based on handheld terminal |
| CN106815580A (en) * | 2016-12-23 | 2017-06-09 | 上海集成电路研发中心有限公司 | A kind of license plate locating method |
| CN107292898A (en) * | 2017-05-04 | 2017-10-24 | 浙江工业大学 | A kind of car plate shadow Detection and minimizing technology based on HSV |
| CN107578044B (en) * | 2017-05-08 | 2020-03-17 | 浙江工业大学 | License plate positioning method based on license plate edge texture features |
| CN107578044A (en) * | 2017-05-08 | 2018-01-12 | 浙江工业大学 | A kind of license plate locating method based on car plate Edge texture feature |
| CN107578379A (en) * | 2017-07-12 | 2018-01-12 | 河海大学 | A kind of processing method of chess robot to checkerboard image |
| CN107609555A (en) * | 2017-09-15 | 2018-01-19 | 北京文安智能技术股份有限公司 | Detection method of license plate, model recognizing method and relevant apparatus using it |
| CN107609555B (en) * | 2017-09-15 | 2020-10-27 | 北京文安智能技术股份有限公司 | License plate detection method, vehicle type identification method applying license plate detection method and related device |
| CN108197521A (en) * | 2017-12-31 | 2018-06-22 | 南京理工大学 | A kind of leggy Quick Response Code obtains identification device and method |
| CN108596172A (en) * | 2018-04-11 | 2018-09-28 | 航天科技控股集团股份有限公司 | A kind of the vehicle binding device and method of car networking client |
| CN110398182A (en) * | 2019-07-04 | 2019-11-01 | 西人马帝言(北京)科技有限公司 | Automatic scoring round target system, method, equipment and medium |
| CN110398182B (en) * | 2019-07-04 | 2022-02-01 | 西人马帝言(北京)科技有限公司 | Automatic target scoring system, method, device and medium |
| CN110569833A (en) * | 2019-09-06 | 2019-12-13 | 上海应用技术大学 | License plate positioning method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105117726B (en) | 2018-10-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105117726A (en) | License plate positioning method based on multi-feature area accumulation | |
| Kong et al. | General road detection from a single image | |
| CN104318258B (en) | Time domain fuzzy and kalman filter-based lane detection method | |
| CN103383733B (en) | A kind of track based on half machine learning video detecting method | |
| CN102999753B (en) | License plate locating method | |
| CN102982304B (en) | Utilize polarized light image to detect the method and system of vehicle location | |
| Li et al. | Nighttime lane markings recognition based on Canny detection and Hough transform | |
| CN108280450A (en) | A kind of express highway pavement detection method based on lane line | |
| CN105493141B (en) | Unstructured road border detection | |
| CN106682586A (en) | Method for real-time lane line detection based on vision under complex lighting conditions | |
| CN105022990A (en) | Water surface target rapid-detection method based on unmanned vessel application | |
| CN105205489A (en) | License plate detection method based on color texture analyzer and machine learning | |
| CN104050450A (en) | Vehicle license plate recognition method based on video | |
| CN102708356A (en) | Automatic license plate positioning and recognition method based on complex background | |
| CN106887004A (en) | A kind of method for detecting lane lines based on Block- matching | |
| CN104809433A (en) | Zebra stripe detection method based on maximum stable region and random sampling | |
| CN102842037A (en) | Method for removing vehicle shadow based on multi-feature fusion | |
| CN110008900B (en) | Method for extracting candidate target from visible light remote sensing image from region to target | |
| CN104766344B (en) | Vehicle checking method based on movement edge extractor | |
| CN106815583A (en) | A kind of vehicle at night license plate locating method being combined based on MSER and SWT | |
| CN103106409A (en) | Composite character extraction method aiming at head shoulder detection | |
| CN104463138A (en) | Text positioning method and system based on visual structure attribute | |
| CN107133588A (en) | Vehicle identification method based on vehicle window feature extraction | |
| CN103679196A (en) | Method for automatically classifying people and vehicles in video surveillance | |
| CN108765456B (en) | Target tracking method and system based on linear edge characteristics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20181009 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |