CN103136514B - A kind of parking event detecting method based on bi-directional tracking - Google Patents
A kind of parking event detecting method based on bi-directional tracking Download PDFInfo
- Publication number
- CN103136514B CN103136514B CN201310045477.7A CN201310045477A CN103136514B CN 103136514 B CN103136514 B CN 103136514B CN 201310045477 A CN201310045477 A CN 201310045477A CN 103136514 B CN103136514 B CN 103136514B
- Authority
- CN
- China
- Prior art keywords
- video image
- frame
- suspicious
- region
- threshold 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.)
- Expired - Fee Related
Links
Abstract
The invention provides a kind of parking event detecting method based on bi-directional tracking, by being divided into multiple pieces of regions, dynamic background extracts, calculate the absolute value sum of pixel difference in present frame each block region corresponding to background, repeat above-mentioned steps to process, being labeled as suspicious piece and assignment is 255, carry out connected domain analysis, obtain the minimum enclosed rectangle of connected domain, mark suspicious region, the movement locus of this suspicious region detected, the running orbit detected is analyzed, determine whether suspicious region is parking.Detection method of the present invention is not by environmental restraint, can detect in real time, and the analysis in conjunction with tracker wire can remove shade, illumination better, leave the impact of thing and surface gathered water, this algorithm can realize detecting accurately in real time road exception parking event, rescue and the process of traffic hazard can be carried out timely and effectively, prevent second accident from occurring, and then ensure the safety that road runs.
Description
Technical field
The invention belongs to field of video detection, be specifically related to a kind of parking event detecting method based on bi-directional tracking.
Background technology
In recent years, along with the continuous increase of the volume of traffic, the generation of traffic hazard is also more and more frequent, in numerous accident, the traffic hazard caused due to parking violation is particularly serious, because Parking has sporadic and randomness, not easily Timeliness coverage and eliminating, once have an accident, traffic jam will be caused, have a strong impact on the normal pass ability of road, and easily cause the generation of follow-up accident, so that lead to very serious consequence.So it is particularly important to set up parking detection system.
At present, automatic traffic event detection algorithm totally can be divided into Indirect Detecting Method and the large class of direct detecting method two.Indirect Detecting Method analyzes by arranging traffic detector collection traffic parameter on road and judge the generation of traffic events, and this method reaction velocity is slow, and reliability is low, and is unfavorable for monitoring, is not the developing direction of following detection method.Direct detecting method is the video by gathering road, then video image processing technology is used to detect traffic abnormity, it is better than Indirect Detecting Method far away in detection speed and reliability, what major part was used in practice is all direct detecting method, and therefore direct detecting method is the developing direction of society traffic event automatic detection system.
Summary of the invention
For shortcomings and deficiencies of the prior art, the object of this invention is to provide a kind of parking event detecting method based on bi-directional tracking, it can detect Parking effectively.
To achieve these goals, the present invention adopts following technical scheme to be achieved:
Based on a parking event detecting method for bi-directional tracking, the method is carried out according to following steps:
Step one, is divided into multiple pieces of regions by each frame of video image, carries out initializing variable setting to each block region of each frame, is each block region and arranges a counter CN and be initialized as 0;
Step 2, carry out dynamic background extraction to the video image in each block region of the first frame, and background video image and original video image are divided into multiple pieces of regions under the same coordinate system, division methods is with step one;
Step 3, if there is stable background, calculate the absolute value sum of pixel difference in present frame each block region corresponding to background, if this value is less than threshold value A, then counter CN has added 1, otherwise counter O reset; If also there is not stable background, then forward step 4 to and continue to perform, wherein:
The span of described threshold value A is 500 ~ 600;
Step 4, from the second frame to p frame, repetition step 2 and step 3 process;
Step 5, when the counter CN of a certain piece reaches threshold value B, is labeled as suspicious piece by this block, and is 255 by the gray-scale value assignment of each pixel in this suspicious piece, otherwise forwards step 4 continuation execution to, wherein:
The span of described threshold value B is 110 ~ 120;
Step 6, carries out connected domain analysis to suspicious piece, obtains the minimum enclosed rectangle of connected domain, if the width of minimum enclosed rectangle is greater than threshold value C, is then suspicious region by this zone marker, otherwise is labeled as non-suspicious region, wherein:
The span of described threshold value C is 5 ~ 10;
Step 7, m frame video image nearest before reading current suspicious region place video image, a frame video image nearest for distance current video image is designated as m frame, a frame video image is farthest designated as the 1st frame, the m-B frame video image that tracking range current video image is far away, if the movement locus that this suspicious region can be detected from described m-B frame video image, then forwards step 8 to and continues to judge, otherwise this suspicious region is no longer processed, wherein:
The span of described m is 190 ~ 200;
Step 8, analyzes the running orbit detected, if the angle of movement locus line and lane line is less than a threshold value D, then forwards step 9 to and continues to judge, otherwise no longer process this suspicious region, wherein:
The span of described threshold value D is 30 ° ~ 40 °;
Step 9, read the n frame video image after the video image of current suspicious region place, follow the tracks of this n frame video image, if the movement locus of this suspicious region can not be detected from described n frame video image, then show that this suspicious region is for stopping, otherwise show that this suspicious region is not stop, detect complete, wherein:
The span of described n is 50 ~ 60.
The present invention detects based on the Parking of bi-directional tracking, not by environmental restraint, can detect in real time, and the analysis in conjunction with tracker wire can remove shade, illumination better, leave the impact of thing and surface gathered water, this algorithm can realize detecting accurately in real time road exception parking event, rescue and the process of traffic hazard can be carried out timely and effectively, prevent second accident from occurring, and then ensure the safety that road runs.
Accompanying drawing explanation
Fig. 1 is the 3170th frame video image, and the vehicle that black box is framed represents the vehicle being about to stop.
Fig. 2 is the 3230th frame video image.
Fig. 3 is the 3355th frame video image, and white portion is the suspicious region detected, black box represents the minimum enclosed rectangle of this suspicious region.
Fig. 4 is the movement locus line for suspicious region traces into, and represents by black lines.
Fig. 5 is the Parking of report.
Below in conjunction with drawings and Examples, content of the present invention is described in further detail.
Embodiment
The present embodiment provides a kind of parking event detecting method based on bi-directional tracking, process in units of block, by following the tracks of suspicious piece, Parking is detected, it should be noted that, image handled in procedure of the present invention be in video along positive seasonal effect in time series first two field picture, the second two field picture, the 3rd two field picture ..., m (m is positive integer) two field picture.
If the size of each frame video image is W*H, the size of each piece is w*h, and wherein W is the pixel of each frame video image horizontal direction, and H is the pixel of each frame video image vertical direction, and w is the width in each piece of region, and h is the height in each piece of region.
The method of the present embodiment specifically adopts following steps to realize:
Step one, is divided into multiple pieces of regions by each frame of video image, then the block areal that each frame can be split is N=(W/w) * (H/h), arranges a counter CN and be initialized as 0 to each block;
Step 2, carry out dynamic background extraction to the video image in each block region of the first frame, and background video image and original video image are divided into multiple pieces of regions under the same coordinate system, division methods is with step one;
Step 3, if there is stable background, calculate the absolute value sum of pixel difference in present frame each block region corresponding to background, if this value is less than threshold value A, then counter CN has added 1, otherwise counter O reset; If also there is not stable background, then forward step 4 to and continue to perform, wherein:
The span of described threshold value A is 500 ~ 600;
Step 4, from the second frame to p frame, repetition step 2 and step 3 process;
Step 5, when the counter CN of a certain piece reaches threshold value B, is labeled as suspicious piece by this block, and is 255 by the gray-scale value assignment of each pixel in this suspicious piece, otherwise forwards step 4 continuation execution to, wherein:
The span of described threshold value B is 110 ~ 120;
Step 6, carries out connected domain analysis to suspicious piece, obtains the minimum enclosed rectangle of connected domain, if the width of minimum enclosed rectangle is greater than threshold value C, is then suspicious region by this zone marker, otherwise is labeled as non-suspicious region, wherein:
The span of described threshold value C is 5 ~ 10;
Step 7, m frame video image nearest before reading current suspicious region place video image, a frame video image nearest for distance current video image is designated as m frame, a frame video image is farthest designated as the 1st frame, the m-B frame video image that tracking range current video image is far away, if the movement locus that this suspicious region can be detected from described m-B frame video image, then forwards step 8 to and continues to judge, otherwise this suspicious region is no longer processed, wherein:
The span of described m is 190 ~ 200;
Step 8, analyzes the running orbit detected, if the angle of movement locus line and lane line is less than a threshold value D, then forwards step 9 to and continues to judge, otherwise no longer process this suspicious region, wherein:
The span of described threshold value D is 30 ° ~ 40 °;
Step 9, read the n frame video image after the video image of current suspicious region place, follow the tracks of this n frame video image, if the movement locus of this suspicious region can not be detected from described n frame video image, then show that this suspicious region is for stopping, otherwise show that this suspicious region is not stop, detect complete, wherein:
The span of described n is 50 ~ 60.
Below provide specific embodiments of the invention, it should be noted that the present invention is not limited to following specific embodiment, all equivalents done on technical scheme basis all fall into protection scope of the present invention.
Embodiment:
As shown in Figures 1 to 5, it is the real-time road video image in section, Xi'an, the sample frequency of this video is that 25 frames are per second, video image size is 720 × 288, the size of every block is 8 × 6, then every frame video image is divided into 90 × 48 blocks, this road vehicles is more, the Target Segmentation threshold value A chosen is 500, and abnormal mass detection threshold B is 120, and vehicle width threshold value C is 8, reading present frame frame number m is forward 200, namely the frame number followed the tracks of forward is 80 frames, and reading present frame frame number n is backward 50, and the video image frame number namely backward followed the tracks of is 50 frames.Defer to said method, above-mentioned video image is processed.
As can be seen from the figure, after the suspicious region in video image being detected, video image is followed the tracks of, 80 frames before video image stops can detect running orbit, and the angle of movement locus and lane line is very little, 50 frames after stopping then do not detect running orbit, so can judge that this suspicious region is as stopping, and carry out Realtime Alerts.
Claims (1)
1. based on a parking event detecting method for bi-directional tracking, it is characterized in that, the method is carried out according to following steps:
Step one, is divided into multiple pieces of regions by each frame of video image, carries out initializing variable setting to each block region of each frame, is each block region and arranges a counter CN and be initialized as 0;
Step 2, carry out dynamic background extraction to the video image in each block region of the first frame, and background video image and original video image are divided into multiple pieces of regions under the same coordinate system, division methods is with step one;
Step 3, if there is stable background, calculate the absolute value sum of pixel difference in present frame each block region corresponding to background, if this value is less than threshold value A, then counter CN has added 1, otherwise counter O reset; If also there is not stable background, then forward step 4 to and continue to perform, wherein:
The span of described threshold value A is 500 ~ 600;
Step 4, from the second frame to p frame, repetition step 2 and step 3 process;
Step 5, when the counter CN of a certain piece reaches threshold value B, is labeled as suspicious piece by this block, and is 255 by the gray-scale value assignment of each pixel in this suspicious piece, otherwise forwards step 4 continuation execution to, wherein:
The span of described threshold value B is 110 ~ 120;
Step 6, carries out connected domain analysis to suspicious piece, obtains the minimum enclosed rectangle of connected domain, if the width of minimum enclosed rectangle is greater than threshold value C, is then suspicious region by this zone marker, otherwise is labeled as non-suspicious region, wherein:
The span of described threshold value C is 5 ~ 10;
Step 7, m frame video image nearest before reading current suspicious region place video image, a frame video image nearest for distance current video image is designated as m frame, a frame video image is farthest designated as the 1st frame, the m-B frame video image that tracking range current video image is far away, if the movement locus that this suspicious region can be detected from described m-B frame video image, then forwards step 8 to and continues to judge, otherwise this suspicious region is no longer processed, wherein:
The span of described m is 190 ~ 200;
Step 8, analyzes the running orbit detected, if the angle of movement locus line and lane line is less than a threshold value D, then forwards step 9 to and continues to judge, otherwise no longer process this suspicious region, wherein:
The span of described threshold value D is 30 ° ~ 40 °;
Step 9, read the n frame video image after the video image of current suspicious region place, follow the tracks of this n frame video image, if the movement locus of this suspicious region can not be detected from described n frame video image, then show that this suspicious region is for stopping, otherwise show that this suspicious region is not stop, detect complete, wherein:
The span of described n is 50 ~ 60.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310045477.7A CN103136514B (en) | 2013-02-05 | 2013-02-05 | A kind of parking event detecting method based on bi-directional tracking |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310045477.7A CN103136514B (en) | 2013-02-05 | 2013-02-05 | A kind of parking event detecting method based on bi-directional tracking |
Publications (2)
Publication Number | Publication Date |
---|---|
CN103136514A CN103136514A (en) | 2013-06-05 |
CN103136514B true CN103136514B (en) | 2016-03-30 |
Family
ID=48496322
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201310045477.7A Expired - Fee Related CN103136514B (en) | 2013-02-05 | 2013-02-05 | A kind of parking event detecting method based on bi-directional tracking |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103136514B (en) |
Families Citing this family (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103886753B (en) * | 2014-03-31 | 2016-09-21 | 北京易华录信息技术股份有限公司 | A kind of signal lamp control crossroad exception parking reason quickly confirms system and method |
CN104574351B (en) * | 2014-08-06 | 2017-07-11 | 深圳市捷顺科技实业股份有限公司 | A kind of method for detecting parking stalls based on Video processing |
KR102203410B1 (en) * | 2014-10-20 | 2021-01-18 | 삼성에스디에스 주식회사 | Method and Apparatus for Setting Region of Interest |
CN104504730B (en) * | 2014-12-19 | 2017-07-11 | 长安大学 | A kind of differentiating method to parking cars and leaving thing |
CN106297278B (en) * | 2015-05-18 | 2019-12-20 | 杭州海康威视数字技术股份有限公司 | Method and system for querying a projectile vehicle |
CN105160326A (en) * | 2015-09-15 | 2015-12-16 | 杭州中威电子股份有限公司 | Automatic highway parking detection method and device |
CN107133610B (en) * | 2017-06-01 | 2020-09-01 | 电子科技大学 | Visual detection and counting method for traffic flow under complex road conditions |
CN113077657B (en) * | 2021-03-30 | 2022-07-05 | 上海华兴数字科技有限公司 | Method and device for alarming safety distance between vehicles |
CN115393413B (en) * | 2022-08-24 | 2023-04-14 | 珠海安士佳电子有限公司 | Intelligent area recognition alarm method and system |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006048238A (en) * | 2004-08-02 | 2006-02-16 | Toshiba Corp | Image processor and image processing program |
CN101183427A (en) * | 2007-12-05 | 2008-05-21 | 浙江工业大学 | Computer vision based peccancy parking detector |
CN102110366A (en) * | 2011-03-28 | 2011-06-29 | 长安大学 | Block-based accumulated expressway vehicle parking event detecting method |
CN102637360A (en) * | 2012-04-01 | 2012-08-15 | 长安大学 | Video-based road parking event detection method |
CN102831773A (en) * | 2011-06-16 | 2012-12-19 | 国民技术股份有限公司 | RFID-based traffic violation lane-changing detection system |
-
2013
- 2013-02-05 CN CN201310045477.7A patent/CN103136514B/en not_active Expired - Fee Related
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2006048238A (en) * | 2004-08-02 | 2006-02-16 | Toshiba Corp | Image processor and image processing program |
CN101183427A (en) * | 2007-12-05 | 2008-05-21 | 浙江工业大学 | Computer vision based peccancy parking detector |
CN102110366A (en) * | 2011-03-28 | 2011-06-29 | 长安大学 | Block-based accumulated expressway vehicle parking event detecting method |
CN102831773A (en) * | 2011-06-16 | 2012-12-19 | 国民技术股份有限公司 | RFID-based traffic violation lane-changing detection system |
CN102637360A (en) * | 2012-04-01 | 2012-08-15 | 长安大学 | Video-based road parking event detection method |
Non-Patent Citations (2)
Title |
---|
基于块计数触发的背景更新;陈伟强;《中国科技信息 》;20111201(第23期);44-45 * |
基于视频跟踪的车辆行为分析技术研究;朱会强;《中国优秀硕士学位论文全文数据库信息科技辑》;20120115(第01期);54 * |
Also Published As
Publication number | Publication date |
---|---|
CN103136514A (en) | 2013-06-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103136514B (en) | A kind of parking event detecting method based on bi-directional tracking | |
US8903133B2 (en) | Periodic stationary object detection system and periodic stationary object detection method | |
JP5997276B2 (en) | Three-dimensional object detection device and foreign object detection device | |
CN103500324B (en) | Violent behavior recognition methods based on video monitoring | |
CN102289940B (en) | Hybrid differential-based traffic flow detection method | |
CN103456024B (en) | A kind of moving target gets over line determination methods | |
CN101739694B (en) | Image analysis-based method and device for ultrahigh detection of high voltage transmission line | |
CN104835323B (en) | Multi-target public transport passenger flow detection method combining with electronic fence | |
CN109598187A (en) | Obstacle recognition method, differentiating obstacle and railcar servomechanism | |
CN102768804A (en) | Video-based traffic information acquisition method | |
CN105744232A (en) | Method for preventing power transmission line from being externally broken through video based on behaviour analysis technology | |
CN101739686A (en) | Moving object tracking method and system thereof | |
CN102568206B (en) | Video monitoring-based method for detecting cars parking against regulations | |
CN103226891B (en) | Video-based vehicle collision accident detection method and system | |
CN103544753B (en) | A kind of banister control method and system | |
CN102622886A (en) | Video-based method for detecting violation lane-changing incident of vehicle | |
CN102968625A (en) | Ship distinguishing and tracking method based on trail | |
CN105160326A (en) | Automatic highway parking detection method and device | |
CN103150549A (en) | Highway tunnel fire detecting method based on smog early-stage motion features | |
CN103236191A (en) | Video-based safety precaution method for vehicle merging from highway ramp | |
CN104866827A (en) | Method for detecting people crossing behavior based on video monitoring platform | |
CN103345840A (en) | Video detection method of road crossing event at cross road | |
CN102110366B (en) | Block-based accumulated expressway vehicle parking event detecting method | |
CN107516423B (en) | Video-based vehicle driving direction detection method | |
CN103150901A (en) | Abnormal traffic condition detection method based on vehicle motion vector field analysis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | 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 | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160330 Termination date: 20220205 |
|
CF01 | Termination of patent right due to non-payment of annual fee |