CN105678312A - Vehicle insurance mark identification method - Google Patents
Vehicle insurance mark identification method Download PDFInfo
- Publication number
- CN105678312A CN105678312A CN201410659005.5A CN201410659005A CN105678312A CN 105678312 A CN105678312 A CN 105678312A CN 201410659005 A CN201410659005 A CN 201410659005A CN 105678312 A CN105678312 A CN 105678312A
- Authority
- CN
- China
- Prior art keywords
- car insurance
- image
- insurance mark
- mark
- car
- 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
Abstract
The present invention discloses a vehicle insurance mark identification method. The method comprises: image capture and obtaining, vehicle insurance mark positioning, image segmentation, measurement normalization, vehicle insurance mark identification, feature matching and result output. The vehicle insurance mark identification may be taken as an important supplement and development for license plate identification, vehicle logo identification and vehicle type identification, may be utilized to criminal case detection of a public security system, and especially may provide an efficient and powerful tool for suspect vehicle checking. The vehicle insurance mark identification is able to save precious time for criminal case detection, save a lot of manpower and material resources and put into more criminal case detection.
Description
Technical field
The present invention relates to a kind of car insurance sign.
Background technology
Automobile is the important vehicles of modern society, and the use of automobile is also more and more frequent. Meanwhile, automobile is also utilized by offender repeatedly as one of guilty tool. At present, Car license recognition, vehicle-logo recognition and vehicle cab recognition technology are widely used in traffic intelligent management system, and in public security system case is detected, the investigation of suspected vehicles are had vital role. But, if suspect is with the use of fake license, deck and blocks the hiding suspected vehicles information of the means such as Che Biao, personnel will be detected to case and the detection works such as the investigation of suspected vehicles are brought inconvenience, not only waste a large amount of human and material resources, the more important thing is the quality time that can incur loss through delay detection case, the stable and compatible of impact society. So, this just need a kind of other householder method and system cracking of cases is carried out necessary supplementing.
Now, the marks (hereinafter referred to as " car insurance mark ") such as the insurance of automobile, monitoring, inspection and/or certification all adopt new label printing product, and the management for the ease of traffic department, this type of mark must be pasted on automobile front windowpane upper right side and must not remove, by the time after must specifying the time limit, old mark can be taken off, new mark of exchangining cards containing all personal details and become sworn brothers. In view of the singularity that car insurance mark this kind is fixing for a long time, using car insurance landmark identification also as an important management tool of intelligent transportation field, the Car license recognition that continues, vehicle-logo recognition, an important supplement of the laggard row vehicle identification of vehicle cab recognition and development can be become.
Summary of the invention
For effectively solving above-mentioned technical barrier, the present invention provides a kind of method of car insurance landmark identification, it is possible to effectively improve the investigation efficiency to suspected vehicles in the detecting of public security system case, meanwhile, is conducive to improving the intelligent management level of vehicle.
This car insurance sign includes the training process of off-line and online recognition process, and described off-line training process and ONLINE RECOGNITION process comprise the steps.
Car insurance mark off-line training process:
S101 collected by hand car insurance mark sample, comprises putting in order between collection vehicle insurance character shape, color and car insurance mark;
The image collected is carried out middle value filtering by S102, removes the noise of image;
Image after S103 centering value filtering carries out gray scale normalized;
Image after gray scale normalized is carried out dimension normalization process by S104;
S105 sample binary conversion treatment obtains car insurance mark template base.
Car insurance mark ONLINE RECOGNITION step:
S201 picture catching and acquisition, obtain the image source containing car insurance mark by image acquisition equipment;
S202 car insurance mark is located, and is specially:
The image caught with get is carried out pre-treatment, pre-treatment comprises filtering process, gray processing process and binary conversion treatment, wherein, adopt the two dimension median filter fast short-cut counting method that coloured image is carried out filtering process, by the time level and smooth image effect, then further smoothed image is carried out gray processing process, obtain gray-scale map picture; Finally adopt global threshold method to carry out binary conversion treatment, obtain the car insurance sign image after two values;
The search of car insurance mark and location, position according to car insurance mark locating information search vehicle insurance mark region from pretreated image, position process, and it carries out dividing processing to car insurance mark region, obtain car insurance mark region image;
S203 Iamge Segmentation, goes out single unit vehicle insurance mark at car insurance mark region Iamge Segmentation;
S204 dimension normalization, the single unit vehicle insurance mark being partitioned into is translated, rotates, contract put, the process of the dimension normalization such as standard cutting, obtain standard-sized single unit vehicle and insure sign image;
S205 car insurance landmark identification, carries out shape, quantity identification to the single unit vehicle insurance mark after segmentation, putting in order between car insurance mark and car insurance mark is identified;
S206 characteristic matching, puts in order the car insurance character shape after identification, quantity or car insurance mark and carries out characteristic matching with the feature templates in car insurance mark template base, draw matching result;
S207 Output rusults.
The present invention, owing to taking above technical scheme, has the following advantages:
1, by adopting technique scheme, can be that public security system is in case detecting, particularly suspected vehicles investigation aspect provides an efficient strong instrument, both for cracking of cases has striven for the quality time, in turn save a large amount of human and material resources, it is possible to put in the detecting of more cases and go;
2, technique scheme is adopted, it is possible to as an important supplement and the development of Car license recognition, vehicle-logo recognition and vehicle cab recognition, for intelligent transportation field provides an important traffic administration instrument.
Accompanying drawing explanation
Fig. 1 is car insurance mark off-line learning process schematic diagram.
Fig. 2 is car insurance mark ONLINE RECOGNITION block diagram.
Embodiment
Below in conjunction with drawings and Examples, the car insurance sign in the present invention is described in detail.
Fig. 1 is the car insurance mark off-line learning process schematic diagram of the embodiment of the present invention. As shown in Figure 1, car insurance mark off-line learning process comprises: S101 collected by hand car insurance mark sample, S102 intermediate value filtering process, S103 gray scale normalized, S104 dimension normalization processes, and S105 sample binary conversion treatment obtains car insurance mark mark template base.
Specifically, S101 collected by hand car insurance mark sample; By staff's collected by hand, completing the classification work of various car insurance mark sample, car insurance mark sample comprises shape, the color of different vehicle insurance mark, and putting in order between car insurance mark, arranges direction.
S102 intermediate value filtering process; Adopting the fast short-cut counting method of two dimension median filter to be processed by the car insurance mark sample collected, eliminate the salt-pepper noise of picture, complete the smoothing processing to picture, algorithm steps is as follows:
Step 1: first detect entrance parameter, carry out image boundary adjustment;
Step 2: distribution output image internal memory district, definition window array, definition histogram number;
Step 3, line by line, first process pixel point of process detection whether every row by column; When being first process pixel point of row, histogram array resets, and reads in the pixel value in window to window array, the histogram of pixel in statistical window, and result, stored in histogram array, returns step 3; When not being first process pixel point of row, former window array first row pixel value is that lower target histogram array unit value subtracts 1 respectively, window is to moving right row, read in the graph data of the rightest column position of new window, it is increased to the right of window array, newly increasing row pixel value is that lower target histogram array unit value adds 1 respectively, returns step 3;
Step 4: entire image process is complete, terminates.
S103 gray scale normalized, the image after the process of centering value filtering carries out gray scale normalized, completes the image to collecting under different light intensity, light source direction and compensates, and weakens the change of the figure image signal caused merely due to illumination variation.
S104 dimension normalization processes, and the translation of image after completing gray scale normalized, rotation, contracting are put, standard cutting etc.
S105 sample binary conversion treatment obtains car insurance mark template base; Finally image is carried out binary conversion treatment, obtain the graphic information of standard, and set up car insurance mark template base.
Car insurance mark off-line learning is mainly carried out prerequisite for car insurance mark ONLINE RECOGNITION and is prepared, and ensures the exactness of car insurance mark ONLINE RECOGNITION.
Fig. 2 is the car insurance mark ONLINE RECOGNITION process schematic diagram of the embodiment of the present invention. As shown in Figure 2, car insurance mark ONLINE RECOGNITION process comprises S201 picture catching and acquisition, and S202 car insurance mark is located, S203 Iamge Segmentation, S204 car insurance landmark identification, S205 characteristic matching, S206 output matching result.
Specifically, S201 picture catching and acquisition, obtain the image source containing car insurance mark by image acquisition equipment.
S202 car insurance mark is located, and is specially:
The image caught with get is carried out pre-treatment, pre-treatment comprises filtering process, gray processing process and binary conversion treatment, wherein, adopt the two dimension median filter fast short-cut counting method that coloured image is carried out filtering process, by the time level and smooth image effect, then further smoothed image is carried out gray processing process, obtain gray-scale map picture; Finally adopt global threshold method to carry out binary conversion treatment, obtain the car insurance sign image after two values;
The search of car insurance mark and location, position according to car insurance mark locating information search vehicle insurance mark region from pretreated image, position process, and it carries out dividing processing to car insurance mark region, obtain car insurance mark region image.
S203 Iamge Segmentation, goes out single unit vehicle insurance mark at car insurance mark region Iamge Segmentation.
S204 dimension normalization, the single unit vehicle insurance mark being partitioned into is translated, rotates, contract put, the process of the dimension normalization such as standard cutting, obtain standard-sized single unit vehicle and insure sign image.
S205 car insurance landmark identification, carries out shape, quantity identification to the single unit vehicle insurance mark after segmentation, putting in order between car insurance mark and car insurance mark is identified.
S206 characteristic matching, puts in order the car insurance character shape after identification, quantity or car insurance mark and carries out characteristic matching with the feature templates in car insurance mark template base, draw matching result.
S207 Output rusults.
Claims (3)
1. a car insurance sign, it is characterised in that: comprise off-line training process and ONLINE RECOGNITION process.
2. car insurance sign according to claim 1, it is characterised in that: car insurance mark off-line learning process comprises collected by hand car insurance mark sample (S101), intermediate value filtering process (S102), gray scale normalized (S103), dimension normalization process (S104) and sample binary conversion treatment and obtains car mark template base (S105);
Collected by hand car insurance mark sample (S103); By staff's collected by hand, completing the classification work of various car insurance mark sample, car insurance mark sample comprises shape, the color of different vehicle insurance mark, and putting in order between car insurance mark, arranges direction;
Intermediate value filtering process (S102); Adopting the fast short-cut counting method of two dimension median filter to be processed by the car insurance mark sample collected, eliminate the salt-pepper noise of picture, complete the smoothing processing to picture, algorithm steps is as follows:
Step 1: first detect entrance parameter, carry out image boundary adjustment;
Step 2: distribution output image internal memory district, definition window array, definition histogram number;
Step 3, line by line, first process pixel point of process detection whether every row by column; When being first process pixel point of row, histogram array resets, and reads in the pixel value in window to window array, the histogram of pixel in statistical window, and result, stored in histogram array, returns step 3; When not being first process pixel point of row, former window array first row pixel value is that lower target histogram array unit value subtracts 1 respectively, window is to moving right row, read in the graph data of the rightest column position of new window, it is increased to the right of window array, newly increasing row pixel value is that lower target histogram array unit value adds 1 respectively, returns step 3;
Step 4: entire image process is complete, terminates;
Gray scale normalized (S103), the image after the process of centering value filtering carries out gray scale normalized, completes the image to collecting under different light intensity, light source direction and compensates, and weakens the change of the figure image signal caused merely due to illumination variation;
Dimension normalization process (S104), the translation of image after completing gray scale normalized, rotation, contracting are put, standard cutting etc.;
Sample binary conversion treatment obtains car insurance mark template base (S105); Finally image is carried out binary conversion treatment, obtain the graphic information of standard, and set up car insurance mark template base.
3. car insurance sign according to claim 1, it is characterised in that: car insurance mark ONLINE RECOGNITION process is:
Picture catching, with obtaining (S201), obtains the image source containing car insurance mark by image acquisition equipment;
Car insurance mark location (S202), is specially:
The image caught with get is carried out pre-treatment, pre-treatment comprises filtering process, gray processing process and binary conversion treatment, wherein, adopt the two dimension median filter fast short-cut counting method that coloured image is carried out filtering process, by the time level and smooth image effect, then further smoothed image is carried out gray processing process, obtain gray-scale map picture; Finally adopt global threshold method to carry out binary conversion treatment, obtain the car insurance sign image after two values;
The search of car insurance mark and location, position according to car insurance mark locating information search vehicle insurance mark region from pretreated image, position process, and it carries out dividing processing to car insurance mark region, obtain car insurance mark region image;
Iamge Segmentation (S203), goes out single unit vehicle insurance mark at car insurance mark region Iamge Segmentation;
Dimension normalization (S204), the single unit vehicle insurance mark being partitioned into is translated, rotates, contract put, the process of the dimension normalization such as standard cutting, obtain standard-sized single unit vehicle and insure sign image;
Car insurance landmark identification (S205), carries out shape, quantity identification to the single unit vehicle insurance mark after segmentation, putting in order between car insurance mark and car insurance mark is identified;
Characteristic matching (S206), puts in order the car insurance character shape after identification, quantity or car insurance mark and carries out characteristic matching with the feature templates in car insurance mark template base, draw matching result;
Output rusults (S207).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410659005.5A CN105678312A (en) | 2014-11-19 | 2014-11-19 | Vehicle insurance mark identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410659005.5A CN105678312A (en) | 2014-11-19 | 2014-11-19 | Vehicle insurance mark identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN105678312A true CN105678312A (en) | 2016-06-15 |
Family
ID=56944748
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410659005.5A Pending CN105678312A (en) | 2014-11-19 | 2014-11-19 | Vehicle insurance mark identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105678312A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874942A (en) * | 2017-01-21 | 2017-06-20 | 江苏大学 | A kind of object module fast construction method semantic based on regular expression |
| CN111401369A (en) * | 2020-03-26 | 2020-07-10 | 上海眼控科技股份有限公司 | Vehicle body danger sign detection method, device, equipment and storage medium |
-
2014
- 2014-11-19 CN CN201410659005.5A patent/CN105678312A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106874942A (en) * | 2017-01-21 | 2017-06-20 | 江苏大学 | A kind of object module fast construction method semantic based on regular expression |
| CN111401369A (en) * | 2020-03-26 | 2020-07-10 | 上海眼控科技股份有限公司 | Vehicle body danger sign detection method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Asphalt pavement pothole detection and segmentation based on wavelet energy field | |
| CN108090423B (en) | Depth license plate detection method based on thermodynamic diagram and key point regression | |
| CN107392103A (en) | The detection method and device of road surface lane line, electronic equipment | |
| CN103512762B (en) | Image processing method, device and train failure detection system | |
| CN108090429B (en) | Vehicle type recognition method for graded front face bayonet | |
| CN102364496B (en) | Method and system for identifying automobile license plates automatically based on image analysis | |
| CN103136528B (en) | A kind of licence plate recognition method based on dual edge detection | |
| CN105405142A (en) | Edge defect detection method and system for glass panel | |
| CN112200172B (en) | Driving region detection method and device | |
| CN103310231A (en) | Auto logo locating and identifying method | |
| CN107909009B (en) | Obstacle detection method and device based on road surface learning | |
| CN102693423A (en) | Method for precise positioning of license plate in strong light conditions | |
| CN103400500A (en) | Vehicle information data acquisition device and method | |
| CN105719283A (en) | Road surface crack image detection method based on Hessian matrix multi-scale filtering | |
| CN106803087A (en) | A kind of car number automatic identification method and system | |
| CN103093201A (en) | Car logo locating and recognizing method and system | |
| CN102346153A (en) | Method for detecting tunnel defect | |
| CN105807335A (en) | Vehicle chassis inspection method and system | |
| CN111126393A (en) | Vehicle appearance refitting judgment method and device, computer equipment and storage medium | |
| CN105913059A (en) | Vehicle VIN code automatic identifying system and control method therefor | |
| CN104834932A (en) | Matlab algorithm of automobile license plate identification | |
| Pandya et al. | Morphology based approach to recognize number plates in India | |
| CN107358236A (en) | A kind of number-plate number identifying system and method based on camera device | |
| WO2022033484A1 (en) | Garbage recognition method and system for intelligent sweeper | |
| Abou Chacra et al. | Municipal infrastructure anomaly and defect detection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20160615 |
|
| WD01 | Invention patent application deemed withdrawn after publication |