基于视频与微波技术融合的高速公路交通事件检测系统研究
|
摘要
随着我国经济的快速发展,高速公路的通车里程、机动车的保有量不断增加,随之而来的交通事故对司乘人员的生命安全及经济发展的影响不断加剧,设计一种快速、准确的交通事件检测系统,提高交通安全势在必行。
基于视频的交通事件自动检测系统和基于微波的交通事件自动检测系统都具有各自的优势与不足,然而,它们的优势存在互补性,如果将这两种检测技术进行融合,就能够得到一种快速、准确、稳定的交通事件自动检测系统。本文设计了一种结合视频技术与微波技术的高速公路交通事件自动检测系统,并对系统中的关键技术进行了深入研究。
首先设计一种基于改进平均法的自适应背景模型,利用背景差分法分割出运动车辆目标,并对目标进行形态学滤波处理,提取运动车辆目标矩形区域,利用基于Kalman滤波预测与Camshift算法相结合的跟踪算法实现了对运动车辆目标的准确跟踪,根据跟踪算法获取的运动车辆目标的质心位置及矩形区域相关位置坐标,设计了一种利用车辆目标之间质心相对距离变化判断车辆碰撞事件及判断拥堵事件的算法。最后针对实验路段以往交通事件视频数据进行分析,得到交通事件发生后交通流变化的规律,结合此规律与冲击波理论,确定了微波交通检测器布设间距与数据采集周期,利用BP神经网络对所采集的数据进行了处理、分析,验证了本文所确定的微波检查器布设间距与数据采集周期的合理性,实现了视频与微波技术的融合。
对比现有的交通事件自动检测系统,本文所设计的结合视频与微波技术的高速公路交通事件自动检测系统具有更高的针对性和实用性,微波交通检测器的布设与数据采集周期设置更加合理,为实际应用打下基础。
With the development of national economy, the traffic mileage of freeway and retain number of automobile are increasing continuously, the influence of passengers'and drive's safety as well as economic development caused by followed traffic incidents is intensifying, it is imperative to design a rapid, accurate automatic incident detection system to promote traffic safety.
Video-based automatic incident detection system and microwave-based automatic incident detection system have advantages and disadvantages respectively, however, their advantages are complementary, a fast, accurate and stable automatic incident detection system will be obtained if combining the two detection techniques together. An automatic incident detection system based video technology and microwave technology combined is designed, and the key technologies of the system in this paper are studied deeply.
An adaptive background model based on improved averaging algorithm is designed firstly, then get the moving vehicles objectives by background subtraction method, and deal with the objectives by morphological filtering, extract the destination rectangle area of moving vehicles, a tracking algorithm combined Kalman filtering with Camshift algorithm is proposed to realize tracking moving vehicles objects accurately, according to the centroids of moving vehicles and the position of rectangular area, vehicles collision judgment algorithm by judging the distance changing of the vehicles centroids and congestion incident judgment algorithm are proposed. Then, analyze the traffic incident video data of experimental road in the past, the variation of traffic flow after the traffic incidents is obtained, then combine this law and shock wave theory, to determine the layout and data acquisition cycle of microwave traffic detector, Processing and analyzing the data collected from the microwave traffic detector by using BP neural network, verify the method of layout and data acquisition cycle of microwave traffic detector in this paper are reasonable, realize the video and microwave technologies fusion。
Comparing the existing automatic incident detection system,the incident detection system based video technology and microwave technology combined has higher pertinence and practicability, the layout and data acquisition cycle of microwave traffic detector are more reasonable,which lays the foundation for practical application.
基于视频的交通事件自动检测系统和基于微波的交通事件自动检测系统都具有各自的优势与不足,然而,它们的优势存在互补性,如果将这两种检测技术进行融合,就能够得到一种快速、准确、稳定的交通事件自动检测系统。本文设计了一种结合视频技术与微波技术的高速公路交通事件自动检测系统,并对系统中的关键技术进行了深入研究。
首先设计一种基于改进平均法的自适应背景模型,利用背景差分法分割出运动车辆目标,并对目标进行形态学滤波处理,提取运动车辆目标矩形区域,利用基于Kalman滤波预测与Camshift算法相结合的跟踪算法实现了对运动车辆目标的准确跟踪,根据跟踪算法获取的运动车辆目标的质心位置及矩形区域相关位置坐标,设计了一种利用车辆目标之间质心相对距离变化判断车辆碰撞事件及判断拥堵事件的算法。最后针对实验路段以往交通事件视频数据进行分析,得到交通事件发生后交通流变化的规律,结合此规律与冲击波理论,确定了微波交通检测器布设间距与数据采集周期,利用BP神经网络对所采集的数据进行了处理、分析,验证了本文所确定的微波检查器布设间距与数据采集周期的合理性,实现了视频与微波技术的融合。
对比现有的交通事件自动检测系统,本文所设计的结合视频与微波技术的高速公路交通事件自动检测系统具有更高的针对性和实用性,微波交通检测器的布设与数据采集周期设置更加合理,为实际应用打下基础。
With the development of national economy, the traffic mileage of freeway and retain number of automobile are increasing continuously, the influence of passengers'and drive's safety as well as economic development caused by followed traffic incidents is intensifying, it is imperative to design a rapid, accurate automatic incident detection system to promote traffic safety.
Video-based automatic incident detection system and microwave-based automatic incident detection system have advantages and disadvantages respectively, however, their advantages are complementary, a fast, accurate and stable automatic incident detection system will be obtained if combining the two detection techniques together. An automatic incident detection system based video technology and microwave technology combined is designed, and the key technologies of the system in this paper are studied deeply.
An adaptive background model based on improved averaging algorithm is designed firstly, then get the moving vehicles objectives by background subtraction method, and deal with the objectives by morphological filtering, extract the destination rectangle area of moving vehicles, a tracking algorithm combined Kalman filtering with Camshift algorithm is proposed to realize tracking moving vehicles objects accurately, according to the centroids of moving vehicles and the position of rectangular area, vehicles collision judgment algorithm by judging the distance changing of the vehicles centroids and congestion incident judgment algorithm are proposed. Then, analyze the traffic incident video data of experimental road in the past, the variation of traffic flow after the traffic incidents is obtained, then combine this law and shock wave theory, to determine the layout and data acquisition cycle of microwave traffic detector, Processing and analyzing the data collected from the microwave traffic detector by using BP neural network, verify the method of layout and data acquisition cycle of microwave traffic detector in this paper are reasonable, realize the video and microwave technologies fusion。
Comparing the existing automatic incident detection system,the incident detection system based video technology and microwave technology combined has higher pertinence and practicability, the layout and data acquisition cycle of microwave traffic detector are more reasonable,which lays the foundation for practical application.
引文
[1]张智勇,朱立伟,高速公路机电系统新技术及应用[M].北京：人民交通出版社,2008.5
[2]刘伟铭.高速公路系统控制方法[M].北京：人民交通出版社,1998
[3]赵志刚,高速公路发展现状及前景分析[J].交通标准化,2009,8：71-73
[4]陈斌,高速公路交通意外事件管理关键技术[M].2007.2
[5]齐志刚,我国高速公路发展的特点分析[J].2009,3／4：109,110
[6]王岩,道路安全事故危险点的判定和成因分析[D].长沙,2003,6：65-67
[7]李相勇,道路交通事故预测方法研究[D].成都,2004,38
[8]中国新闻网,2009年全国道路交通事故238351起 致67759人死[EB／OL].
http://www.chinanews.com.cn/gn/news/2010/01-09/2063204.shtml,2010,1,9
[9]赵祥模,靳引利,张洋.高速公路监控系统理论及应用[M].北京,电子工业出版社,2003
[10]Reducing the severity of road injuries through post impact care[C]. Brussels, European Transport Safety Council, Post Impact Care Working Party,1999
[11]陈斌,金炜东,四类交通事故检测算法的分析[J].交通科技与经济,2005,3：50-52
[12]张敬磊,交通事件检测算法研究进展[J].武汉理工大学学报,2005,2：215-218
[13]徐学才,刘澜,自动事件检测算法的比较评估[J].交通科技与经济,2004,2：42.44
[14]姜紫峰,刘小坤,基于神经网络的交通事件检测算法[J].西安公路交通大学学报,2000,3：67-69／73
[15]Geneidy A M,Ahrned M, Bertin, Toward valiardation of freeway loop detector speed measurements using transit probe data[C].7th International IEEE Conference Proceedings on Intelligent Transportation Systems,2004:779-784
[16]Zhang X P, Nihan N L, Wang Y H, mproved dual-loop detection system for collecting real-time truck data[J].Transportation Research Record,2005:109-115
[17]Weng G Q, He T N, Chen M J, The vehicle's classification recognition system based on DTW algorithm[C].5th World Congress Proceedings on Intellignt Control and Automation,2004:4169-4171
[18]王长峰,微波车辆检测器在河南高速公路中的应用[J].中国交通信息产业，2006,2：123-124
[19]Holden E J, Owens R, Segmenting Occluded Objects Using a Motion Snake[C].Asian Conference on Computer vision(ACCV),2004:SIV-3
[20]王孝通,基于光流法的舰船运动要素测定原理研究[J].中国航海,2004,2：7-10
[21]梁国山,朱秀昌,用于运动人体检测的改进的帧差法[J].视频应用与工程,2009,33：133-135
[22]于万霞,杜太行,基于彩色空间的背景帧差法视频车辆检测[J].计算机仿真.2010,1：285-287/308
[23]张建荣,冀小平,裴亮,基于背景差法的视频图像分割[J].中国新技术产品,2009,1：1
[24]杜丽丽,智能交通系统中视频车辆检测技术的研究[J].中国海洋大学,2009,6：32
[25]骆迪,基于视频技术的车辆违章行为检测[D].西华大学.2008,1：21
[26]Sekim, Fujiwarah, Sumik, A Robust Background Subtraction Method for Changing Background[C]. Proceedings of IEEE Workshop on Applications of Computer Vision,2000:207-213
[27]张建荣,冀小平,裴亮,基于背景差法的视频图像分割[J]中国新技术新产品,2009,1：1
[28]阮秋琦,数字图像处理学[M].北京：电子工业出版社,1999：429-449
[29]黄建清,李中益,张利珍.基于视频车辆目标检测的阴影去除的研究[J].计算机与信息技术,2007，102(5)：77-79
[30]Shoichi Araki, Takashi Matsuoaka, Naokazu Yokoya, et al.Real-time tracking of multiple moving object contours in a moving camera image sequence[C].IEICE Trans Inf&Syst,2000, E83-D(7)
[31]J.Shi, and C.Tomasi. Good Feature to track[C].Proceedings of IEEE Conference on Computer Vision and pattern Recognition.1991:593-600
[32]A.K.Jain, YZhong, and S.Lakshmanan, Object Matching Using Deformable Templates[J].IEEE Trans. Pattern Analysis and Machine Intelligence,1996(3):267-278
[33]M.Bertalmio, GS apiro, and QRandall, Morphing Active Contours. Proc. Int'l Conf.Scale-Space Theories in Computer Vision,1999:46-57
[34]Fukunage K, Hostetler L D. The estimation of the gradient of a density function with application in pattern recogrition[J]. IEEE Transactions of InformationTheory.1975,21(1):32-40
[35]Comaniciu D, Meer P. Mean shift:A robust application toward feature space analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619
[36]Comaniciu D, Ramesh V, Meer P.Real-time tracking of non-rigid objects using mean shift[A].IEEE Conference on Computer Vision and Pattern Recognition[C].2000,2:142-149
[37]Allen John G.Xu Richard Y D.J in Jesse S. Object tracking using Camshift algorithm and multiple quantized feature spaces[C]. Pan-Sydney Area Workshop on Visual Information Processing VDP2003, Sydney, Australia,2003:1-5
[38]Boyle M. The Effects of Capture Conditions on the CAMSHIFT Face Tracker [R].Alberta, Canada: Department of Computer Science, University of Calgary,2001:45-47
[39]Allan JG, Xur Y D, Jin JS. Object tracking using CamShift algorithm and multiple quantized feature spaces[D].Sydney:University of Sydney,2006:3-7
[40]黄建新.Kalman滤波的人体运动位置跟踪算法[J].华侨大学学报,2003,24(3)：254-256
[41]张江山,朱光嘉.一种基于Kalman滤波的视频对象跟踪方法[J].中国图象图形学报,2002,7(A)：606-609
[42]王长峰,微波车辆检测器在河南高速公路中的应用[J].中国交通信息产业.2006.2：123-24
[43]彭春华,刘建业,刘岳峰,等,车辆检测传感器综述[J].2007.6：4-7／11
[44]向敬成,张明友,雷达系统[M].电子工业出版社2001
[45]微波交通检测器的设置.中国人民共和国国建标准.
[46]张琦.动态OD估计的交通监测器优化布设研究[D].吉林：吉林大学,2005
[47]伍建国,王峰.城市道路交通数据采集系统检测器优化布点研究[J].公路交通科技,2004,2,21(2)：88-92
[48]刘伟铭.高速公路系统控制方法[M].北京：人民交通出版社,1998
[49]张汝华.高速道路交通流信息采集与处理理论及方法研究[D].上海：同济大学,2005
[50]李国勇,智能控制及其MATLAB实现[M].北京,电子工业出版社2007.4
[2]刘伟铭.高速公路系统控制方法[M].北京：人民交通出版社,1998
[3]赵志刚,高速公路发展现状及前景分析[J].交通标准化,2009,8：71-73
[4]陈斌,高速公路交通意外事件管理关键技术[M].2007.2
[5]齐志刚,我国高速公路发展的特点分析[J].2009,3／4：109,110
[6]王岩,道路安全事故危险点的判定和成因分析[D].长沙,2003,6：65-67
[7]李相勇,道路交通事故预测方法研究[D].成都,2004,38
[8]中国新闻网,2009年全国道路交通事故238351起 致67759人死[EB／OL].
http://www.chinanews.com.cn/gn/news/2010/01-09/2063204.shtml,2010,1,9
[9]赵祥模,靳引利,张洋.高速公路监控系统理论及应用[M].北京,电子工业出版社,2003
[10]Reducing the severity of road injuries through post impact care[C]. Brussels, European Transport Safety Council, Post Impact Care Working Party,1999
[11]陈斌,金炜东,四类交通事故检测算法的分析[J].交通科技与经济,2005,3：50-52
[12]张敬磊,交通事件检测算法研究进展[J].武汉理工大学学报,2005,2：215-218
[13]徐学才,刘澜,自动事件检测算法的比较评估[J].交通科技与经济,2004,2：42.44
[14]姜紫峰,刘小坤,基于神经网络的交通事件检测算法[J].西安公路交通大学学报,2000,3：67-69／73
[15]Geneidy A M,Ahrned M, Bertin, Toward valiardation of freeway loop detector speed measurements using transit probe data[C].7th International IEEE Conference Proceedings on Intelligent Transportation Systems,2004:779-784
[16]Zhang X P, Nihan N L, Wang Y H, mproved dual-loop detection system for collecting real-time truck data[J].Transportation Research Record,2005:109-115
[17]Weng G Q, He T N, Chen M J, The vehicle's classification recognition system based on DTW algorithm[C].5th World Congress Proceedings on Intellignt Control and Automation,2004:4169-4171
[18]王长峰,微波车辆检测器在河南高速公路中的应用[J].中国交通信息产业，2006,2：123-124
[19]Holden E J, Owens R, Segmenting Occluded Objects Using a Motion Snake[C].Asian Conference on Computer vision(ACCV),2004:SIV-3
[20]王孝通,基于光流法的舰船运动要素测定原理研究[J].中国航海,2004,2：7-10
[21]梁国山,朱秀昌,用于运动人体检测的改进的帧差法[J].视频应用与工程,2009,33：133-135
[22]于万霞,杜太行,基于彩色空间的背景帧差法视频车辆检测[J].计算机仿真.2010,1：285-287/308
[23]张建荣,冀小平,裴亮,基于背景差法的视频图像分割[J].中国新技术产品,2009,1：1
[24]杜丽丽,智能交通系统中视频车辆检测技术的研究[J].中国海洋大学,2009,6：32
[25]骆迪,基于视频技术的车辆违章行为检测[D].西华大学.2008,1：21
[26]Sekim, Fujiwarah, Sumik, A Robust Background Subtraction Method for Changing Background[C]. Proceedings of IEEE Workshop on Applications of Computer Vision,2000:207-213
[27]张建荣,冀小平,裴亮,基于背景差法的视频图像分割[J]中国新技术新产品,2009,1：1
[28]阮秋琦,数字图像处理学[M].北京：电子工业出版社,1999：429-449
[29]黄建清,李中益,张利珍.基于视频车辆目标检测的阴影去除的研究[J].计算机与信息技术,2007，102(5)：77-79
[30]Shoichi Araki, Takashi Matsuoaka, Naokazu Yokoya, et al.Real-time tracking of multiple moving object contours in a moving camera image sequence[C].IEICE Trans Inf&Syst,2000, E83-D(7)
[31]J.Shi, and C.Tomasi. Good Feature to track[C].Proceedings of IEEE Conference on Computer Vision and pattern Recognition.1991:593-600
[32]A.K.Jain, YZhong, and S.Lakshmanan, Object Matching Using Deformable Templates[J].IEEE Trans. Pattern Analysis and Machine Intelligence,1996(3):267-278
[33]M.Bertalmio, GS apiro, and QRandall, Morphing Active Contours. Proc. Int'l Conf.Scale-Space Theories in Computer Vision,1999:46-57
[34]Fukunage K, Hostetler L D. The estimation of the gradient of a density function with application in pattern recogrition[J]. IEEE Transactions of InformationTheory.1975,21(1):32-40
[35]Comaniciu D, Meer P. Mean shift:A robust application toward feature space analysis[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2002,24(5):603-619
[36]Comaniciu D, Ramesh V, Meer P.Real-time tracking of non-rigid objects using mean shift[A].IEEE Conference on Computer Vision and Pattern Recognition[C].2000,2:142-149
[37]Allen John G.Xu Richard Y D.J in Jesse S. Object tracking using Camshift algorithm and multiple quantized feature spaces[C]. Pan-Sydney Area Workshop on Visual Information Processing VDP2003, Sydney, Australia,2003:1-5
[38]Boyle M. The Effects of Capture Conditions on the CAMSHIFT Face Tracker [R].Alberta, Canada: Department of Computer Science, University of Calgary,2001:45-47
[39]Allan JG, Xur Y D, Jin JS. Object tracking using CamShift algorithm and multiple quantized feature spaces[D].Sydney:University of Sydney,2006:3-7
[40]黄建新.Kalman滤波的人体运动位置跟踪算法[J].华侨大学学报,2003,24(3)：254-256
[41]张江山,朱光嘉.一种基于Kalman滤波的视频对象跟踪方法[J].中国图象图形学报,2002,7(A)：606-609
[42]王长峰,微波车辆检测器在河南高速公路中的应用[J].中国交通信息产业.2006.2：123-24
[43]彭春华,刘建业,刘岳峰,等,车辆检测传感器综述[J].2007.6：4-7／11
[44]向敬成,张明友,雷达系统[M].电子工业出版社2001
[45]微波交通检测器的设置.中国人民共和国国建标准.
[46]张琦.动态OD估计的交通监测器优化布设研究[D].吉林：吉林大学,2005
[47]伍建国,王峰.城市道路交通数据采集系统检测器优化布点研究[J].公路交通科技,2004,2,21(2)：88-92
[48]刘伟铭.高速公路系统控制方法[M].北京：人民交通出版社,1998
[49]张汝华.高速道路交通流信息采集与处理理论及方法研究[D].上海：同济大学,2005
[50]李国勇,智能控制及其MATLAB实现[M].北京,电子工业出版社2007.4