城市道路车辆违章系统的设计与实现
|
摘要
随着交通运输的飞速发展,智能交通系统在现代交通系统中发挥着越来越大的作用。交通系统中最难以解决的是城市交通中的交通堵塞现象。目前,交通堵塞现象很大一部分原因是由于车主在繁忙的道路上违章停车造成的,针对这个问题,交管部门在违停区安装了球形摄像机,通过人工的方式控制摄像机对违停的车辆进行抓拍,事后对录像进行剪辑,并将它们作为处罚的依据。在公安部应用创新计划的支持下,我们开发了一套车辆违停自动抓拍系统,该系统能够实现对违停区内车辆的检测与跟踪,判断车辆是否有违停的行为,进而控制摄像机对违停的车辆进行抓拍,最后在后台软件的协同下完成录像的剪辑。
本系统采用TMS320DM642DSP芯片作为开发平台。在DSP/BIOS提供的功能模块的基础上完成了本系统下位机的编码,并通过CCS集成开发环境实现了系统的调试及优化。为了实现系统的网络通信功能,我们采用NDK开发套件实现了本系统的TCP/IP协议栈,并按照系统的需求对其进行了裁剪和优化。
在对违停区内车辆的检测过程中,系统将摄像机采集到的帧图像通过TVP5150BPS解码芯片解码后生成YUV(4：2：2)格式的数字信号,经过Y分量的提取和4*4的均值压缩后,DSP对图像数据进行处理。系统采用背景差分法实现对运动目标的检测。背景生成采用了统计平均法,并用surendra算法对背景进行自适应更新。最后利用OTSU算法对前景图像进行自适应阈值分割,得到包含运动目标信息的二值图像。
对车辆违停行为的判断,本文采用的是基于违停规则的车辆跟踪算法。系统在初始化时建立了一个以每台车辆信息为结点的链表,而当前帧中处在违停区内的车辆信息保存在栈中,通过栈中元素与链表中每个结点的匹配,从而实现链表的插入,删除和结点数据的更新,最后对链表中结点的信息进行判断,可以判断出车辆是否有违停的行为。
为了实现系统的配置管理、违章视频的剪切和系统的升级功能,我们开发了能与下位机进行网络通信的上位机软件。通过这些软件,工作人员能够方便地对违停的车辆进行处理,该系统能够充分满足交管部门的需求。
As the development of transportation, the intelligent transportation system is playing a more and more important role in the modern transportation system. Recently, the traffic jams are the most bothering things in our daily life which is caused by illegal parking most of times. To solve this problem, the traffic control department has equipped cameras at the no parking area, staffs can capture vehicles which stop in the no parking area for a short of time by controlling these cameras. After this, they have to find out illegal parked vehicles by clipping videos and thus punish vehicle owners. Under the support of the innovative program of the Military of Public Security, we research and developed the illegal parked automatic capturing system, which can detect and trace vehicles in no parking area. If these are illegal parked vehicles, the system can automatic control camera and capture these cars. By running software on PC, we can get all illegal parked vehicles information.
We choose TMS320DM642DSP as the system's hardware platform. According to varies modules supplied by DSP/BIOS, we accomplished embed device's code. Later, we debug and optimize the code by CCS integrated development environment, we focused on how to realize the TCP/IP stack of TMS320DM642and how to optimize it. In order to communicate with software on PC by network, we realized the embed device's TCP/IP stack by the kit of NDK and optimize it according to the system's function need.
In the process of vehicle detection, images captured by the camera is analog signal, which can be changed to YUV(4:2:2) digital signal format by a decoder chip TVP5150BPS, the formatted digital signal can be directly inputted to TMS320DM642's video port. Then, we only extract the Y component from the YUV format and compress it by4*4mean compression. We choose background subtraction method as the moving detection algorithm, the background is generated by method of average and updated adaptively by algorithm of surendra. After that, the foreground image is segmented adaptively by OTSU algorithm. So we can get a binary image which contains all moving targets information.
By determine vehicles'behavior in no parking area, we choose a vehicle tracking method based on illegal parked rules. The system maintains a linked list, whose each node contains a vehicle's state information. Each time we get a frame, we compare vehicles'information that extracted from this fame with the linked list's each node. By doing this, we can do insert, delete operation to the linked list and update each node's information. After all this, we only need to traverse the linked list to determine each vehicle's behavior in no parking area.
In order to manage and configure all embedded devices, we developed some software running on PC, by which, working staffs can deal with illegal events easily. The system can fully meet the requirement of Military of Public Security.
本系统采用TMS320DM642DSP芯片作为开发平台。在DSP/BIOS提供的功能模块的基础上完成了本系统下位机的编码,并通过CCS集成开发环境实现了系统的调试及优化。为了实现系统的网络通信功能,我们采用NDK开发套件实现了本系统的TCP/IP协议栈,并按照系统的需求对其进行了裁剪和优化。
在对违停区内车辆的检测过程中,系统将摄像机采集到的帧图像通过TVP5150BPS解码芯片解码后生成YUV(4：2：2)格式的数字信号,经过Y分量的提取和4*4的均值压缩后,DSP对图像数据进行处理。系统采用背景差分法实现对运动目标的检测。背景生成采用了统计平均法,并用surendra算法对背景进行自适应更新。最后利用OTSU算法对前景图像进行自适应阈值分割,得到包含运动目标信息的二值图像。
对车辆违停行为的判断,本文采用的是基于违停规则的车辆跟踪算法。系统在初始化时建立了一个以每台车辆信息为结点的链表,而当前帧中处在违停区内的车辆信息保存在栈中,通过栈中元素与链表中每个结点的匹配,从而实现链表的插入,删除和结点数据的更新,最后对链表中结点的信息进行判断,可以判断出车辆是否有违停的行为。
为了实现系统的配置管理、违章视频的剪切和系统的升级功能,我们开发了能与下位机进行网络通信的上位机软件。通过这些软件,工作人员能够方便地对违停的车辆进行处理,该系统能够充分满足交管部门的需求。
As the development of transportation, the intelligent transportation system is playing a more and more important role in the modern transportation system. Recently, the traffic jams are the most bothering things in our daily life which is caused by illegal parking most of times. To solve this problem, the traffic control department has equipped cameras at the no parking area, staffs can capture vehicles which stop in the no parking area for a short of time by controlling these cameras. After this, they have to find out illegal parked vehicles by clipping videos and thus punish vehicle owners. Under the support of the innovative program of the Military of Public Security, we research and developed the illegal parked automatic capturing system, which can detect and trace vehicles in no parking area. If these are illegal parked vehicles, the system can automatic control camera and capture these cars. By running software on PC, we can get all illegal parked vehicles information.
We choose TMS320DM642DSP as the system's hardware platform. According to varies modules supplied by DSP/BIOS, we accomplished embed device's code. Later, we debug and optimize the code by CCS integrated development environment, we focused on how to realize the TCP/IP stack of TMS320DM642and how to optimize it. In order to communicate with software on PC by network, we realized the embed device's TCP/IP stack by the kit of NDK and optimize it according to the system's function need.
In the process of vehicle detection, images captured by the camera is analog signal, which can be changed to YUV(4:2:2) digital signal format by a decoder chip TVP5150BPS, the formatted digital signal can be directly inputted to TMS320DM642's video port. Then, we only extract the Y component from the YUV format and compress it by4*4mean compression. We choose background subtraction method as the moving detection algorithm, the background is generated by method of average and updated adaptively by algorithm of surendra. After that, the foreground image is segmented adaptively by OTSU algorithm. So we can get a binary image which contains all moving targets information.
By determine vehicles'behavior in no parking area, we choose a vehicle tracking method based on illegal parked rules. The system maintains a linked list, whose each node contains a vehicle's state information. Each time we get a frame, we compare vehicles'information that extracted from this fame with the linked list's each node. By doing this, we can do insert, delete operation to the linked list and update each node's information. After all this, we only need to traverse the linked list to determine each vehicle's behavior in no parking area.
In order to manage and configure all embedded devices, we developed some software running on PC, by which, working staffs can deal with illegal events easily. The system can fully meet the requirement of Military of Public Security.
引文
[1]靳茜文.北京市智能交通管理系统的安全效益定量评价研究[D].北京：北京邮电大学,2011.
[2]刘开国.重庆城市轨道交通调度指挥系统的网络化模式[J].城市轨道交通研究,2011(02)：109-112.
[3]傅宇辉.城市智能交通信息发布系统的关键技术研究[D].上海：上海交通大学,2008.
[4]赵克栋.视频监控系统设计与工程应用[D].北京：北京邮电大学,2009.
[5]钱毅.基于3G标准的嵌入式网络视频监控系统设计与实现[D].厦门：厦门大学,2008.
[6]王卫岳.基于TMS320DM642的车牌识别系统[D].杭州：浙江大学,2007.
[7]高鹏.基于嵌入式的监测监控系统硬件开发与实现[D].北京：北京邮电大学,2007.
[8]李天长.基于DM642的嵌入式实时图像处理的研究[D].重庆：重庆大学,2008.
[9]Texas Instruments. TMS320C6000 CPU and Instruction Set Reference Guide (Rev. F) [M]. Literature Number:SPRU189F October 2000.
[10]傅海东.基于DSP和FPGA的车牌识别系统设计及实现[D].成都：电子科技大学,2009.
[11]邓锦豪.基于视频图像的行人检测算法研究[D].广州：华南理工大学,2011.
[12]张志付.基于背景减法的运动检测算法研究[D].上海：上海交通大学,2008.
[13]Hirai J,Yamaguchi T,Harada H. Extraction of moving object based on fast optical flow estimation[C]. In:Proceedings of ICCAS-SICE 2009. Fukuoka:[s.n],2009:2691-2695.
[14]何卫华,李平,文玉梅,叶波.复杂背景下基于图像融合的运动目标轮廓提取算法[J].计算机应用,2006,26(1)：123-126.
[15]林坤.视频监控中运动人体跟踪与行为分析研究[D].天津：天津大学,2010.
[16]陈锡.视频监控中运动目标检测跟踪技术[D].大连：大连理工大学,2010.
[17]Sebastian Brutzer, Benjamin Hoferlin,Gunther Heidemann.Evaluation of Background Subtraction Techniques for Video Surveillance[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.2011:1937-1944.
[18]Hirai J,Yamaguchi T,Harada H. Extraction of moving object based on fast optical flow estimation[C]. In:Proceedings of ICCAS-SICE 2009. Fukuoka:[s.n],2009:2691-2695.
[19]周建锋.基于光流的运动目标检测方法研究[D].哈尔滨：哈尔滨工业大学,2009.
[20]Xiaojing Song, Lakmal D. Seneviratne, Kaspar Althoefer. A Kalman filter-integrated optical flow method for velocity sensing of mobile robots[J]. IEEE/ASME Transactions on Mechatronics.2011(16):551-563.
[21]Lagae Lieven,Ceulemans Berten, Van Huffel Sabine, Vanrumste Bart. Automatic video detection of body movement during sleep based on optical flow in pediatric patients with epilepsy[J]. Medical and Biological Engineering and Computing.2010(48),923-931.
[22]董颖.基于光流场的视频运动检测[D].济南：山东大学,2008.
[23]胡以静,李政访,跃明.基于光流的运动分析理论及应用[J].2007,15(2)：219-221.
[24]杨建伟.融合MeanShift和卡尔曼滤波的运动目标跟踪算法研究[D].哈尔滨：哈尔滨工业大学,2010.
[25]Takemura, Hiroshi, Mizoguchi, Hiroshi. Robust tracking method by MeanShift using Spatiograms[J]. Proceedings of the SICE Annual Conference.2010:1985-1988.
[26]何革.基于MeanShift算法的视频目标跟踪研究[D].重庆：重庆大学,2010.
[27]Einhaus Julian, Hahn Markus, Wohler Christian, Kummert, Franz. Vehicle tracking and motion prediction in complex urban scenarios[J]. IEEE Intelligent Vehicles Symposium, Proceedings.2010:26-33.
[28]Ma, Kai-Kuang. A state-space super-resolution approach for video reconstruction[J]. Signal, Image and Video Processing[J].2009(3):217-240.
[29]冯仕民.智能视频监控中运动目标跟踪技术研究[D].成都：电子科技大学,2010.
[30]秦晓敏.智能视频监控中运动目标检测与跟踪技术的研究.成都：西南交通大学,2011.
[31]项崇明.基于DM642的网络摄像机研制[D].杭州：浙江大学,2007.
[32]王艳艳.基于DSP的人工视觉图像采集系统[D].上海：上海交通大学,2007.
[33]Texas Instruments. TMS320C6000 TCP/IP Network Developer's Kit[M].2004-11-21:25-26.
[34]罗莉.基于DSP的视频采集压缩系统的硬件设计与实现[D].武汉：华中科技大学,2007.
[35]李建科.基于DSP的车辆目标检测系统设计与研究[D].重庆：重庆大学,2009.
[36]朱晓鼎,张东,刘发志.基于TMS320DM642芯片视频系统的设计与应用[J].计算机工程与设计,2008,29(9)：2233-2238.
[37]缪国远.基于视频的运动目标检测与识别技术的研究[D].大连：大连理工大学,2010,
[38]邬大鹏,程卫平,于盛林.基于帧间差分和运动估计的Camshift目标跟踪算法[J].光电工程,2010,37(1)：55-59.
[39]胡以静,李政访,胡跃明.基于光流的运动分析理论及应用[J].计算机测量与控制,2007：：219-221.
[40]Wali Ali, Alimi Adel M. Event detection from video surveillance data based on optical flow histogram and high-level feature extraction[C]. Linz, Austria:Institute of Electrical and Electronics Engineers Inc,2009:221-225.
[41]侯宏录,李宁鸟,刘迪迪,陈杰.智能视频监控中运动目标检测的研究[J].2012,22(2)：49-52.
[42]Cheong W.L, Mohamaddan S, Kamaruddin, A.M.N. Abg, Yassin A., Case K, Motion detection using periodic background estimation subtraction method[C]. Kuching, Sarawak, Malaysia:IEEE Computer Society,445 Hoes Lane-P.O.Box 1331, Piscataway, NJ 08855-1331, United States,2011.
[43]张文岳.基于视频的机动车跨道违章监测系统[D].天津：天津大学,2009.
[44]史志龙,陈一民,徐杰.视频监控中人的运动检测方法研究与实现[J].2008,44(19)：206-210.
[45]冈萨雷斯.数字图像处理(第二版)[M].阮秋琦,阮宇智.北京：电子工业出版社,2007：496-508.
[46]Chang, ChinShu Chen, Ji Ding, Lin Jui-Sheng, Liao Teh-Lu. Design and implementation of real-time object tracking system using the gaussian motion model and the Otsu Algorithm[C]. United States:IEEE Computer Society,2009:140-143.
[2]刘开国.重庆城市轨道交通调度指挥系统的网络化模式[J].城市轨道交通研究,2011(02)：109-112.
[3]傅宇辉.城市智能交通信息发布系统的关键技术研究[D].上海：上海交通大学,2008.
[4]赵克栋.视频监控系统设计与工程应用[D].北京：北京邮电大学,2009.
[5]钱毅.基于3G标准的嵌入式网络视频监控系统设计与实现[D].厦门：厦门大学,2008.
[6]王卫岳.基于TMS320DM642的车牌识别系统[D].杭州：浙江大学,2007.
[7]高鹏.基于嵌入式的监测监控系统硬件开发与实现[D].北京：北京邮电大学,2007.
[8]李天长.基于DM642的嵌入式实时图像处理的研究[D].重庆：重庆大学,2008.
[9]Texas Instruments. TMS320C6000 CPU and Instruction Set Reference Guide (Rev. F) [M]. Literature Number:SPRU189F October 2000.
[10]傅海东.基于DSP和FPGA的车牌识别系统设计及实现[D].成都：电子科技大学,2009.
[11]邓锦豪.基于视频图像的行人检测算法研究[D].广州：华南理工大学,2011.
[12]张志付.基于背景减法的运动检测算法研究[D].上海：上海交通大学,2008.
[13]Hirai J,Yamaguchi T,Harada H. Extraction of moving object based on fast optical flow estimation[C]. In:Proceedings of ICCAS-SICE 2009. Fukuoka:[s.n],2009:2691-2695.
[14]何卫华,李平,文玉梅,叶波.复杂背景下基于图像融合的运动目标轮廓提取算法[J].计算机应用,2006,26(1)：123-126.
[15]林坤.视频监控中运动人体跟踪与行为分析研究[D].天津：天津大学,2010.
[16]陈锡.视频监控中运动目标检测跟踪技术[D].大连：大连理工大学,2010.
[17]Sebastian Brutzer, Benjamin Hoferlin,Gunther Heidemann.Evaluation of Background Subtraction Techniques for Video Surveillance[C]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition.2011:1937-1944.
[18]Hirai J,Yamaguchi T,Harada H. Extraction of moving object based on fast optical flow estimation[C]. In:Proceedings of ICCAS-SICE 2009. Fukuoka:[s.n],2009:2691-2695.
[19]周建锋.基于光流的运动目标检测方法研究[D].哈尔滨：哈尔滨工业大学,2009.
[20]Xiaojing Song, Lakmal D. Seneviratne, Kaspar Althoefer. A Kalman filter-integrated optical flow method for velocity sensing of mobile robots[J]. IEEE/ASME Transactions on Mechatronics.2011(16):551-563.
[21]Lagae Lieven,Ceulemans Berten, Van Huffel Sabine, Vanrumste Bart. Automatic video detection of body movement during sleep based on optical flow in pediatric patients with epilepsy[J]. Medical and Biological Engineering and Computing.2010(48),923-931.
[22]董颖.基于光流场的视频运动检测[D].济南：山东大学,2008.
[23]胡以静,李政访,跃明.基于光流的运动分析理论及应用[J].2007,15(2)：219-221.
[24]杨建伟.融合MeanShift和卡尔曼滤波的运动目标跟踪算法研究[D].哈尔滨：哈尔滨工业大学,2010.
[25]Takemura, Hiroshi, Mizoguchi, Hiroshi. Robust tracking method by MeanShift using Spatiograms[J]. Proceedings of the SICE Annual Conference.2010:1985-1988.
[26]何革.基于MeanShift算法的视频目标跟踪研究[D].重庆：重庆大学,2010.
[27]Einhaus Julian, Hahn Markus, Wohler Christian, Kummert, Franz. Vehicle tracking and motion prediction in complex urban scenarios[J]. IEEE Intelligent Vehicles Symposium, Proceedings.2010:26-33.
[28]Ma, Kai-Kuang. A state-space super-resolution approach for video reconstruction[J]. Signal, Image and Video Processing[J].2009(3):217-240.
[29]冯仕民.智能视频监控中运动目标跟踪技术研究[D].成都：电子科技大学,2010.
[30]秦晓敏.智能视频监控中运动目标检测与跟踪技术的研究.成都：西南交通大学,2011.
[31]项崇明.基于DM642的网络摄像机研制[D].杭州：浙江大学,2007.
[32]王艳艳.基于DSP的人工视觉图像采集系统[D].上海：上海交通大学,2007.
[33]Texas Instruments. TMS320C6000 TCP/IP Network Developer's Kit[M].2004-11-21:25-26.
[34]罗莉.基于DSP的视频采集压缩系统的硬件设计与实现[D].武汉：华中科技大学,2007.
[35]李建科.基于DSP的车辆目标检测系统设计与研究[D].重庆：重庆大学,2009.
[36]朱晓鼎,张东,刘发志.基于TMS320DM642芯片视频系统的设计与应用[J].计算机工程与设计,2008,29(9)：2233-2238.
[37]缪国远.基于视频的运动目标检测与识别技术的研究[D].大连：大连理工大学,2010,
[38]邬大鹏,程卫平,于盛林.基于帧间差分和运动估计的Camshift目标跟踪算法[J].光电工程,2010,37(1)：55-59.
[39]胡以静,李政访,胡跃明.基于光流的运动分析理论及应用[J].计算机测量与控制,2007：：219-221.
[40]Wali Ali, Alimi Adel M. Event detection from video surveillance data based on optical flow histogram and high-level feature extraction[C]. Linz, Austria:Institute of Electrical and Electronics Engineers Inc,2009:221-225.
[41]侯宏录,李宁鸟,刘迪迪,陈杰.智能视频监控中运动目标检测的研究[J].2012,22(2)：49-52.
[42]Cheong W.L, Mohamaddan S, Kamaruddin, A.M.N. Abg, Yassin A., Case K, Motion detection using periodic background estimation subtraction method[C]. Kuching, Sarawak, Malaysia:IEEE Computer Society,445 Hoes Lane-P.O.Box 1331, Piscataway, NJ 08855-1331, United States,2011.
[43]张文岳.基于视频的机动车跨道违章监测系统[D].天津：天津大学,2009.
[44]史志龙,陈一民,徐杰.视频监控中人的运动检测方法研究与实现[J].2008,44(19)：206-210.
[45]冈萨雷斯.数字图像处理(第二版)[M].阮秋琦,阮宇智.北京：电子工业出版社,2007：496-508.
[46]Chang, ChinShu Chen, Ji Ding, Lin Jui-Sheng, Liao Teh-Lu. Design and implementation of real-time object tracking system using the gaussian motion model and the Otsu Algorithm[C]. United States:IEEE Computer Society,2009:140-143.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |