CN103646544B - Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus - Google Patents
Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus Download PDFInfo
- Publication number
- CN103646544B CN103646544B CN201310567228.4A CN201310567228A CN103646544B CN 103646544 B CN103646544 B CN 103646544B CN 201310567228 A CN201310567228 A CN 201310567228A CN 103646544 B CN103646544 B CN 103646544B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- cloud terrace
- behavioural analysis
- utilize
- recognition methods
- 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.)
- Active
Links
Landscapes
- Image Analysis (AREA)
- Studio Devices (AREA)
Abstract
The invention discloses a kind of vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus, comprise the steps: to utilize the mixed Gaussian background modeling method based on the weak edge of image to carry out sport foreground detection; Utilize the vehicle detecting algorithm based on cascade LBP spatial histogram feature to detect sport foreground region, judge whether this region has vehicle, and accurate positioned vehicle position; The rapid vehicle matching algorithm based on class Fern feature is utilized to mate localizing objects vehicle for a long time; Utilize clarification of objective number of matches to determine whether parking offense, after cradle head preset positions switches, carry out analyzing to target vehicle in advance and identify; Use embedded integrated hardware configuration, realize the repertoires such as the collection of real time video image, vehicle behavioural analysis, The Cloud Terrace interlock, compression of images and Internet Transmission, also there is the equipment Management Functions such as time segment is enable, point presetting bit is enable simultaneously, as long as rear end platform software receives picture and does data preparation screening operation.
Description
Technical field
The invention belongs to technical field of intelligent traffic, relate to a kind of vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus.
Background technology
In intelligent transportation industry, CCTV camera and The Cloud Terrace are very universal, but to the behavioral value of vehicle and the intelligence degree of analysis lower, automatically detect violation event (parking offense, drive in the wrong direction) indifferent, most special messenger of group checks, so not only needs to drop into a large amount of human and material resources, and due to the energy of people and notice limited, under the working environment of high strength, often there is careless omission, thus, can not react abnormal conditions in time.
Summary of the invention
The present invention is exactly to solve above-mentioned the problems of the prior art, provides a kind of vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus.
In order to achieve the above object, the present invention adopts following technical scheme:
Vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus of the present invention, comprises the steps:
The mixed Gaussian background modeling method based on the weak edge of image is utilized to carry out sport foreground detection;
Utilizing based on cascade LBP(localbinarypatterns, local binary patterns feature) vehicle detecting algorithm of spatial histogram feature detects sport foreground region, judges whether this region has vehicle, and accurate positioned vehicle position;
The rapid vehicle matching algorithm based on class fern (Fern) feature is utilized to mate localizing objects vehicle for a long time;
Utilize clarification of objective number of matches to determine whether parking offense, cradle head preset positions is not detecting Static Detection target vehicle under sport foreground after switching, and carries out analyzing and identify target vehicle in advance;
Use embedded integrated hardware configuration, realize the collection of real time video image, vehicle behavioural analysis, The Cloud Terrace interlock, the focusing of camera zoom, vehicle snapshot, Car license recognition, compression of images and Internet Transmission.
Wherein, the mixed Gaussian background modeling method based on the weak edge of image is utilized to carry out the concrete steps of sport foreground detection as follows: first to utilize the large scale SOBEL(Sobel of improvement) operator carries out vertical and horizontal direction convolution and gets maximal value obtaining outline map to image; Then set up mixed Gauss model with Image boundaries and set up background and extraction prospect.
Detecting in sport foreground the step cascade whether having vehicle to exist employs two kinds of sorters to utilize the vehicle detecting algorithm based on cascade LBP spatial histogram feature to confirm: detect and filter out the SVM(supportvectormachine of non-vehicle fast, support vector base) linear classifier and carry out the accurate SVMHIK(Histogramintersectionkernel confirmed, histogram intersection core) Nonlinear Classifier.
Rapid vehicle matching algorithm based on class Fern feature is divided into subregion matching image, utilizes integrogram method to calculate average gray in all subregions as matching criterior.
Utilize clarification of objective number of matches to determine whether parking offense, every frame image is detected the buffer zone that the position of moving target, size and distance put into regular length constantly upgrades, and the up-to-date target signature that detects and historical data traversal are carried out coupling numeration, reaching predetermined value is just parking.
After cradle head preset positions switches, adopt based on the stationary vehicle detection technique of demarcating yardstick more, target vehicle is in advance carried out analyzing and identified, draws near and calculate different vehicle calibration yardsticks.
Use embedded integrated hardware configuration, complete monitoring, intellectual analysis function by four modules such as intelligent analysis module, video acquisition buffer module, The Cloud Terrace main control module and video compress in The Cloud Terrace front end.
The advantage that the present invention has and good effect are:
Vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus of the present invention, real-time examination and analysb can be carried out to target vehicle, candid photograph image is carried out to vehicles peccancy and records a video, record whole process violating the regulations, operating personnel only need to inquire about violating the regulations record, substantially increase the work efficiency of operating personnel.The method can be reacted and record abnormal conditions in time, avoids owing to manually causing careless omission.
Accompanying drawing explanation
Fig. 1 is the computing method of LBP operator of the present invention;
Fig. 2 is LBP spatial histogram of the present invention;
Fig. 3 is the vehicle detecting algorithm process flow diagram based on cascade LBP spatial histogram feature;
Fig. 4 is the block plan of the rapid vehicle matching algorithm based on class Fern feature;
Fig. 5 is system hardware architecture block diagram of the present invention;
Fig. 6 is the process flow diagram of vehicles peccancy detection, vehicle snapshot, car plate analysis identification.
Embodiment
Below in conjunction with the drawings and specific embodiments, the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus of the present invention is described further.Following each embodiment is not only limitation of the present invention for illustration of the present invention.
Vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus of the present invention, comprises the steps:
The mixed Gaussian background modeling method based on the weak edge of image is utilized to carry out sport foreground detection;
Utilize the vehicle detecting algorithm based on cascade LBP spatial histogram feature to detect sport foreground region, judge whether this region has vehicle, and accurate positioned vehicle position;
The rapid vehicle matching algorithm based on class Fern feature is utilized to mate localizing objects vehicle for a long time;
Utilize clarification of objective number of matches to determine whether parking offense, cradle head preset positions is not detecting Static Detection target vehicle under sport foreground after switching, and carries out analyzing and identify target vehicle in advance;
Use embedded integrated hardware configuration, realize the repertoires such as the collection of real time video image, vehicle behavioural analysis, The Cloud Terrace interlock, the focusing of camera zoom, vehicle snapshot, Car license recognition, compression of images and Internet Transmission, also there is the equipment Management Functions such as time segment is enable, point presetting bit is enable simultaneously.
Wherein, utilize that to carry out the concrete steps of sport foreground detection based on the mixed Gaussian background modeling method at the weak edge of image as follows: first utilize the large scale SOBEL operator of improvement to carry out vertical and horizontal direction convolution and get maximal value obtaining outline map to image, more weak edge can be made to strengthen, then set up mixed Gauss model with Image boundaries and set up background and extraction prospect, the illegal vehicle sailing setting regions into can be extracted fast, substantially increase the anti-interference of system and the accuracy of detection.
The vehicle detecting algorithm based on cascade LBP spatial histogram feature is utilized to confirm whether have vehicle to exist in detection sport foreground, this step cascade employs SVM linear classifier and SVMHIK Nonlinear Classifier two kinds of sorters, linear classifier detects fast and filters out non-vehicle, and Nonlinear Classifier confirms accurately.
The computing method of LBP operator as shown in Figure 1, compare for as each pixel of central point and the pixel in surrounding 8 UNICOM territory, if be greater than or equal to, what obtain corresponding position is encoded to 1, otherwise be 0, the coding of these 8 corresponding positions coupled together the decimal representation that obtains 8 bits in a certain order and is the LBP value of this central point pixel.
LBP spatial histogram as shown in Figure 2, vehicle sample is divided into certain cell, every 2x2 cell is designated as a block, LBP histogram is added up in each block, each block in the horizontal direction, the stepping of vertical direction movement is a cell, until travel through whole sample image, and the histogram of each block is together in series obtains the LBP spatial histogram feature of whole sample.
Based on cascade LBP spatial histogram feature vehicle detecting algorithm process flow diagram as shown in Figure 3.LBP spatial histogram feature is extracted to foreground image, utilizes two sorters to carry out classification judgement respectively to each feature, all think have car then to think there is car this position, otherwise think that this position is without car.
Rapid vehicle matching algorithm based on class Fern feature can mate target vehicle for a long time, this algorithm is divided into subregion matching image, integrogram method is utilized to calculate average gray in all subregions as matching criterior, this algorithm travelling speed is fast, to block and illumination variation has certain robustness.
Rapid vehicle matching algorithm based on class Fern feature is divided into 13 sub regions matching image, as shown in Figure 4, respectively add up subregion average mate.
Utilize clarification of objective number of matches to determine whether parking offense, every frame image is detected the buffer zone that the position of moving target, size and distance put into regular length constantly upgrades, and the up-to-date target signature that detects and historical data traversal are carried out coupling count, reach predetermined value and just think parking, all Parameter adjustable joints, substantially increase the adaptability of system.
After cradle head preset positions switches, adopt based on the stationary vehicle detection technique of demarcating yardstick more, can target vehicle in advance be carried out analyzing and be identified.This technology is demarcated yardstick to vehicle detection and is carried out logic optimization, draw near and calculate different vehicle calibration yardsticks, the vehicle of different distance in scene, different size all accurately can be detected by the vehicle detecting algorithm based on cascade LBP spatial histogram feature.
Use embedded integrated hardware configuration, as shown in Figure 5, hardware circuit mainly comprises intellectual analysis main control module, video acquisition buffer module, The Cloud Terrace main control module and video compressing module.By intelligent analysis module, video acquisition buffer module, The Cloud Terrace main control module and video compressing module, monitoring, intellectual analysis function can be completed in The Cloud Terrace front end, this system has fast response time, and take rear end platform resource few, intellectual analysis is without advantages such as time delays.
The repertoires such as the collection of video image, vehicle behavioural analysis, The Cloud Terrace interlock, the focusing of camera zoom, vehicle snapshot, Car license recognition, compression of images and Internet Transmission are realized in The Cloud Terrace front end, also there is the equipment Management Functions such as time segment is enable, point presetting bit is enable simultaneously, as long as rear end platform software receives picture and does data preparation screening operation.
Vehicles peccancy detection, vehicle snapshot, Car license recognition thread realization flow figure are as shown in Figure 6, when there being vehicles peccancy, need panorama sketch be captured, calculate parking duration, after exceeding regulation duration, control The Cloud Terrace, capture feature, carry out Car license recognition, then control The Cloud Terrace, return monitoring position, again capture panoramic pictures.
Claims (7)
1., based on a vehicle behavioural analysis recognition methods for The Cloud Terrace and camera apparatus, it is characterized in that, the method comprises the steps:
The mixed Gaussian background modeling method based on the weak edge of image is utilized to carry out sport foreground detection;
Utilize the vehicle detecting algorithm based on the LBP spatial histogram feature of cascade to detect sport foreground region, judge whether this region has vehicle, and accurate positioned vehicle position;
The rapid vehicle matching algorithm based on class fern feature is utilized to mate localizing objects vehicle for a long time;
Utilize clarification of objective number of matches to determine whether parking offense, cradle head preset positions is not detecting Static Detection target vehicle under sport foreground after switching, and carries out analyzing and identify target vehicle in advance;
Use embedded integrated hardware configuration, realize the collection of real time video image, vehicle behavioural analysis, The Cloud Terrace interlock, the focusing of camera zoom, vehicle snapshot, Car license recognition, compression of images and Internet Transmission.
2. the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus according to claim 1, it is characterized in that, utilize the mixed Gaussian background modeling method based on the weak edge of image to carry out the concrete steps of sport foreground detection as follows: first utilize the large scale SOBEL operator of improvement to carry out vertical and horizontal direction convolution and get maximal value obtaining outline map to image; Then set up mixed Gauss model with Image boundaries and set up background and extraction prospect.
3. the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus according to claim 1, it is characterized in that, detecting the step cascade whether having vehicle to exist in sport foreground employs two kinds of sorters to utilize the vehicle detecting algorithm based on cascade LBP spatial histogram feature to confirm: detect fast and filter out the SVM linear classifier of non-vehicle and carry out the accurate Nonlinear Classifier based on SVMHIK confirmed.
4. the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus according to claim 1, it is characterized in that, rapid vehicle matching algorithm based on class Fern feature is divided into subregion matching image, utilizes integrogram method to calculate average gray in all subregions as matching criterior.
5. the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus according to claim 1, it is characterized in that: utilize clarification of objective number of matches to determine whether parking offense, every frame image is detected the buffer zone that the position of moving target, size and distance put into regular length constantly upgrades, and the up-to-date target signature that detects and historical data traversal are carried out coupling numeration, reaching predetermined value is just parking offense.
6. the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus according to claim 1, it is characterized in that: after cradle head preset positions switches, adopt based on the stationary vehicle detection technique of demarcating yardstick more, carry out analyzing to target vehicle in advance and identify, drawing near and calculate different vehicle calibration yardsticks.
7. the vehicle behavioural analysis recognition methods based on The Cloud Terrace and camera apparatus according to claim 1, it is characterized in that: use embedded integrated hardware configuration, complete monitoring, intellectual analysis function by intelligent analysis module, video acquisition buffer module, The Cloud Terrace main control module and video compress four module in The Cloud Terrace front end.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310567228.4A CN103646544B (en) | 2013-11-15 | 2013-11-15 | Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310567228.4A CN103646544B (en) | 2013-11-15 | 2013-11-15 | Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103646544A CN103646544A (en) | 2014-03-19 |
CN103646544B true CN103646544B (en) | 2016-03-09 |
Family
ID=50251750
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310567228.4A Active CN103646544B (en) | 2013-11-15 | 2013-11-15 | Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103646544B (en) |
Families Citing this family (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104157143B (en) * | 2014-08-15 | 2016-06-01 | 青岛比特信息技术有限公司 | Parking offense detection system and detection method thereof |
CN104778842A (en) * | 2015-04-29 | 2015-07-15 | 深圳市保千里电子有限公司 | Cloud vehicle running track tracing method and system based on vehicle license plate recognition |
CN106446892A (en) * | 2016-09-13 | 2017-02-22 | 东软集团股份有限公司 | Vehicle license plate recognizing method, apparatus and charging pile |
CN106878674B (en) * | 2017-01-10 | 2019-08-30 | 哈尔滨工业大学深圳研究生院 | A kind of parking detection method and device based on monitor video |
CN109035805B (en) * | 2018-09-21 | 2021-03-02 | 深圳市九洲电器有限公司 | Intelligent recognition method and device for license plate number |
CN111291748B (en) * | 2020-01-15 | 2020-12-11 | 广州玖峰信息科技有限公司 | Cascade distributed artificial intelligence case number identification system |
JP7331769B2 (en) * | 2020-04-30 | 2023-08-23 | トヨタ自動車株式会社 | Position estimation system and position estimation method |
CN112183350B (en) * | 2020-09-28 | 2023-07-14 | 天地伟业技术有限公司 | Video-based illegal parking detection method |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1984236A (en) * | 2005-12-14 | 2007-06-20 | 浙江工业大学 | Method for collecting characteristics in telecommunication flow information video detection |
CN101183427A (en) * | 2007-12-05 | 2008-05-21 | 浙江工业大学 | Computer vision based peccancy parking detector |
EP1978470A1 (en) * | 2007-04-05 | 2008-10-08 | Mitsubishi Electric Corporation | Method for detecting objects left-behind in a scene |
JP2012038318A (en) * | 2010-08-10 | 2012-02-23 | Fujitsu Ltd | Target detection method and device |
CN102637257A (en) * | 2012-03-22 | 2012-08-15 | 北京尚易德科技有限公司 | Video-based detection and recognition system and method of vehicles |
CN103035013A (en) * | 2013-01-08 | 2013-04-10 | 东北师范大学 | Accurate moving shadow detection method based on multi-feature fusion |
-
2013
- 2013-11-15 CN CN201310567228.4A patent/CN103646544B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN1984236A (en) * | 2005-12-14 | 2007-06-20 | 浙江工业大学 | Method for collecting characteristics in telecommunication flow information video detection |
EP1978470A1 (en) * | 2007-04-05 | 2008-10-08 | Mitsubishi Electric Corporation | Method for detecting objects left-behind in a scene |
CN101183427A (en) * | 2007-12-05 | 2008-05-21 | 浙江工业大学 | Computer vision based peccancy parking detector |
JP2012038318A (en) * | 2010-08-10 | 2012-02-23 | Fujitsu Ltd | Target detection method and device |
CN102637257A (en) * | 2012-03-22 | 2012-08-15 | 北京尚易德科技有限公司 | Video-based detection and recognition system and method of vehicles |
CN103035013A (en) * | 2013-01-08 | 2013-04-10 | 东北师范大学 | Accurate moving shadow detection method based on multi-feature fusion |
Non-Patent Citations (1)
Title |
---|
基于视频的运动车辆检测和跟踪算法研究;郝士新;《中国优秀硕士学位论文全文数据库 信息科技辑 》;20100715(第7期);I140-433:正文全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN103646544A (en) | 2014-03-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103646544B (en) | Based on the vehicle behavioural analysis recognition methods of The Cloud Terrace and camera apparatus | |
CN104751634B (en) | The integrated application method of freeway tunnel driving image acquisition information | |
CN103927878B (en) | A kind of automatic shooting device for parking offense and automatically grasp shoot method | |
CN104951775B (en) | Railway highway level crossing signal region security intelligent identification Method based on video technique | |
CN103093249B (en) | A kind of taxi identification method based on HD video and system | |
Jadhav et al. | Smart traffic control system using image processing | |
CN102759347B (en) | Online in-process quality control device and method for high-speed rail contact networks and composed high-speed rail contact network detection system thereof | |
CN106571039A (en) | Automatic snapshot system for highway traffic offence | |
CN101789177B (en) | Device and method for detecting and tracking vehicles crossing and pressing the yellow line and for capturing vehicle information | |
CN109949579A (en) | A kind of illegal automatic auditing method that makes a dash across the red light based on deep learning | |
CN108269407B (en) | Security robot capable of automatically managing people stream and logistics | |
CN104464290A (en) | Road traffic parameter collecting and rule violation snapshot system based on embedded double-core chip | |
CN103279756A (en) | Vehicle detecting analysis system and detecting analysis method thereof based on integrated classifier | |
CN101231786A (en) | Vehicle checking method based on video image characteristic | |
CN101937614A (en) | Plug and play comprehensive traffic detection system | |
WO2010077316A1 (en) | Multiple object speed tracking system | |
CN104537360A (en) | Method and system for detecting vehicle violation of not giving way | |
Nodado et al. | Intelligent traffic light system using computer vision with android monitoring and control | |
CN103021179B (en) | Based on the Safe belt detection method in real-time monitor video | |
CN113947731A (en) | Foreign matter identification method and system based on contact net safety inspection | |
CN111723706A (en) | Box type freight car door opening monitoring device and system based on raspberry group | |
Ua-Areemitr et al. | Low-cost road traffic state estimation system using time-spatial image processing | |
CN207938184U (en) | A kind of vehicle-mounted act of violating regulations capturing system | |
CN106960193A (en) | A kind of lane detection apparatus and method | |
CN204375161U (en) | Parking offense automatic snapshot system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C14 | Grant of patent or utility model | ||
GR01 | Patent grant | ||
CP03 | Change of name, title or address | ||
CP03 | Change of name, title or address |
Address after: 300384 Tianjin City Huayuan Industrial Zone (outer ring road No. 8) two Haitai branch Patentee after: Tiandi Weiye Technology Co., Ltd. Address before: 300384 Tianjin city Xiqing District Huayuan new technology Industrial Park (outer ring road No. 8) two Haitai branch Patentee before: Tianjin Tiandy Digital Technology Co., Ltd. |