基于视频监控的车速检测算法研究
|
摘要
随着社会经济的快速发展,交通系统日益复杂。为了给交通系统的管理提供各种实时交通信息,以方便、高效地利用和管理现有的交通系统,基于视频图像处理的交通信息采集系统成为研究的一个热点。本课题作为交通信息采集系统的一个重要子系统,它以实时的车辆行驶速度为研究检测对象,其目的是改进实时车辆速度的采集手段,为道路交通管理提供更为有效的依据,提高交通管理的自动化、现代化水平。
本文致于研究一种灵活、可靠的视频交通流测检测系统。车辆的检测基于车道,在每个车道可以设置一条虚拟检测线和一个虚拟检测区域来检测交通流的车速,虚拟检测线的作用类似于电磁感应线圈。系统通过对视频虚拟检测线的预处理将二维的数字图像信号转化为一维的检测信号。该方法的特点是只需对虚拟检测线区域内的图像进行处理,处理运算的图像区域小,因此减少了运算量,降低了运算负荷。
本文还提出了一种从频率域出发,估计出运动车辆在空间域上位移的算法。利用傅里叶变换的自配准性质和位移特性,用极坐标形式下连续两帧车辆图像相位谱的差可直接估计运动车辆的位移,进而估计出车速u=s/△t。由于本算法分辩能力不低于一个像素,因而可用于车速的高精度测量。
利用本文提出的车速测量算法,对晴天和阴天两种典型天气条件下的交通流视频进行了实验,实验结果表明具有较高的车流量和车速检测精度。
With the rapid development of society and economy,the traffic system becomes more and more complex.In order to provide all kinds of real-time traffic information to the management of traffic system for achieving efficient management and making full use of the traffic system,the study on real-time traffic information collection system based on video image processing has been becoming a focus. As an important subsystem of real-time traffic information collection system, the paper makes the running speed of vehicles as the studying and detecting object and mainly studies on the detecting algorithms of real-time speed and other related ones. The purpose of this subject is to improve the detecting methods of real-time speed and provide more efficient references to increase the automatic and modernizing traffic management levels.
This paper focuses on developing a flexible and reliable system to detect the real-time speed of the traffic flow parameters through image sequences. The vehicle detection is based on the roadway, which sets a virtual line and a virtual-loop area, like the inductive loop sensor, to detect the traffic flow parameters. The key idea of the system is converting the 2-dimensional digital image to 1-dimensional detection signal by virtual detection line preprocessing, which reduces computation load. The trait of this method is only processing little image area within virtual line or virtual-loop and avoiding vehicle tracking in 2-dimensional image, hence the time cost of calculation and the computation burthen is recuced.
An algorithm which is used to estimate the translation of a moving object in spatial domain according to its Fourier transform spectrum is also proposed in the paper. Based on shift theorem of Fourier transform and auto-registration, the algorithm directly estimates the translation with the phase spectrum difference between continuous images of a moving object in polar coordinate system. Vehicle velocity v=s/△t (where s is the displacement of vehicle and At is the interval between two continuous images) is calculated corresponding. It can give the displacement of a moving target with a resolution less than 1 pixel, hence it can be used to measure vehicle velocity with high precision.
Using the algorithms put forwarded in this paper, the traffic flow video files in fine day and cloudy day are experimented and the experimental results show that the detection accuracy is rather high.
本文致于研究一种灵活、可靠的视频交通流测检测系统。车辆的检测基于车道,在每个车道可以设置一条虚拟检测线和一个虚拟检测区域来检测交通流的车速,虚拟检测线的作用类似于电磁感应线圈。系统通过对视频虚拟检测线的预处理将二维的数字图像信号转化为一维的检测信号。该方法的特点是只需对虚拟检测线区域内的图像进行处理,处理运算的图像区域小,因此减少了运算量,降低了运算负荷。
本文还提出了一种从频率域出发,估计出运动车辆在空间域上位移的算法。利用傅里叶变换的自配准性质和位移特性,用极坐标形式下连续两帧车辆图像相位谱的差可直接估计运动车辆的位移,进而估计出车速u=s/△t。由于本算法分辩能力不低于一个像素,因而可用于车速的高精度测量。
利用本文提出的车速测量算法,对晴天和阴天两种典型天气条件下的交通流视频进行了实验,实验结果表明具有较高的车流量和车速检测精度。
With the rapid development of society and economy,the traffic system becomes more and more complex.In order to provide all kinds of real-time traffic information to the management of traffic system for achieving efficient management and making full use of the traffic system,the study on real-time traffic information collection system based on video image processing has been becoming a focus. As an important subsystem of real-time traffic information collection system, the paper makes the running speed of vehicles as the studying and detecting object and mainly studies on the detecting algorithms of real-time speed and other related ones. The purpose of this subject is to improve the detecting methods of real-time speed and provide more efficient references to increase the automatic and modernizing traffic management levels.
This paper focuses on developing a flexible and reliable system to detect the real-time speed of the traffic flow parameters through image sequences. The vehicle detection is based on the roadway, which sets a virtual line and a virtual-loop area, like the inductive loop sensor, to detect the traffic flow parameters. The key idea of the system is converting the 2-dimensional digital image to 1-dimensional detection signal by virtual detection line preprocessing, which reduces computation load. The trait of this method is only processing little image area within virtual line or virtual-loop and avoiding vehicle tracking in 2-dimensional image, hence the time cost of calculation and the computation burthen is recuced.
An algorithm which is used to estimate the translation of a moving object in spatial domain according to its Fourier transform spectrum is also proposed in the paper. Based on shift theorem of Fourier transform and auto-registration, the algorithm directly estimates the translation with the phase spectrum difference between continuous images of a moving object in polar coordinate system. Vehicle velocity v=s/△t (where s is the displacement of vehicle and At is the interval between two continuous images) is calculated corresponding. It can give the displacement of a moving target with a resolution less than 1 pixel, hence it can be used to measure vehicle velocity with high precision.
Using the algorithms put forwarded in this paper, the traffic flow video files in fine day and cloudy day are experimented and the experimental results show that the detection accuracy is rather high.
引文
[1]Kumar P,Sengupta K,Lee A.A comparative study of different color spaces for foreground and shadow detection for traffic monitoring system.The IEEE5~(th)international conference on intelligent transportation systems,Singapore,2002:100-105.
[2]Tseng B.L.,Lin C.Y.,Smith J.R..Real-time video surveillance for traffic monitoring using virtual line analysis.2002 IEEE international coference,2002,2:541-544.
[3]Liu Y.,Payeur P.,Vision-based detection of activity for traffic control.IEEE CCECE 2003,Canadian,2003,2:1347-1350.
[4]Cucchiara R.,Piccardi M.,Prati A..Detecting moving objects,ghosts,and shadows in video streams.IEEE transactions on pattern analysis and machine intelligence.2003,(25):1337-1342.
[5]Yu M.,Jiang CRY.,Yu B.K..An integrative method for video based traffic parameter extraction in ITS.The 2000 IEEE Asia-Pacific Conference,2000:136-139.
[6]Wang J.M.,Chung Y.C.,Chang C.L.etal.Shadow detection and removal for traffic images.Proceeding of the 2004 IEEE international conference on Networking,sensing&control,Taipei,Taiwan,2004:649-654.
[7]杨佩昆.智能交通[M].上海:同济大学出版社,2002.
[8]《中国智能运输系统体系框架》专题组.中国智能运输系统体系框架[M].北京:人民邮电出版社,2003.
[9]史新宏,蔡伯根,穆建成.智能交通系统的发展[J].北方交通大学学报,2002,26(1):29-34.
[10]Rafael C.GONZALEZ,Richard E.WOODS.数字图像处理(第二版)[M].北京:电子工业出版社,2003.
[11]刘智勇.智能交通控制理论及其应用[M].北京:科学出版社,2003.
[12]阮秋琦等译.数字图像处理[M].北京:电子工业出版社,2005.
[13]章毓晋.图像处理和分析[M].北京:清华大学出版社,2000.
[14]向世明.Visual C++数字图像与图形处理[M].北京:电子工业出版社,2003.
[15]刘玮玮.Visual C++视频/音频开发实用工程案例精选[M].北京:人民邮电出版社,2005.
[16]徐慧等.Visual C++数字图像实用工程案例精选[M].北京:人民邮电出版社,2005.
[17]杨枝灵,王开.Visual C++数字图像获取、处理及实践应用[M].北京:人民邮电出版社,2003.
[18]胡小锋,赵辉.Visual C++/MATLAB图像处理与识别应用案例精选[M].北京:人民邮电出版社,2003.
[19]陈纯.计算机图像处理技术与算法[M].北京:清华大学出版社,2003.
[20]马骏.高速公路行车安全距离的分析与研究[J].西安公路交通大学学报,1998,18(4):90-94.
[21]Zhigang zhu,Guangyou Xu,Bo Yang,etal.A real-time vision system for automatic traffic monitoring.Image and Vision Computing,2000,18(20):781-794.
[22]张玲,叶海炳,何伟.一种基于边缘信息的改进车辆检测方法[J].重庆大学学报(自然科学版),2001,27(11):56-58.
[23]郁梅,蒋刚毅等.基于路面标记的车辆检测和计数[J].仪器仪表学报,2002,23(4):386-390.
[24]魏武,张起森等.基于计算机视觉和图像处理的交通流参数检测[J].信息与控制,2004,30(3):257-261.
[25]武想益,郑依展等.视频交通流检测演示系统[J].计算机工程,2003,29(21):151-153.
[26]王夏黎,周明全等.交通流视频检测系统的设计与实现[J].西北大学学报(自然科学版),2004,26(12):207-210.
[27]彭仁明,贺春林等.基于视频的车流量检测(自然科学版)[J].西华师范大学学报,2004,25(4):404-408.
[28]巨永锋,朱辉等.基于计算机视觉的车流量检测算法[J].长安大学学报(自然科学版),2004,24(1):92-95.
[29]王栓等.基于差分图像的多运动目标的检测与跟踪[J].中国图像图形学报,1999,4(6):470-475.
[30]张玲,陈丽敏等.基于视频的改进帧差法在车流量检测中的应用[J].重庆大学学报(自然科学版),2004,27(5):31-33.
[31]吴天昊,于前洋.用傅里叶相移特性估计位移[J].光电工程,2005,32(8):73-76.
[32]王东峰,邹谋炎.傅氏变换的自配准性质及其在纹理识别和图像分割中的应用[J].中国图象图形学报,2003,8(2):140-146.
[33]尹朝征.基于视频虚拟线圈的交通流参数检测[D].清华大学硕士学位论文.2002.
[34]李绪龙.基于图像处理技术的视频交通流信息采集系统研究[D].长安大学硕士学位论文.2003.
[35]王强.基于计算机视觉的交通流参数采集理论与方法研究[D].吉林大学硕士学位论文.2001.
[36]杜楠.复杂气候条件下的视频交通流量检测算法[D].哈尔滨工业大学硕士学位论文.2004.
[37]夏文龙.基于计算机视觉的汽车流量检测研究[D].四川大学硕士学位论文.2003.
[38]童剑军,邹明福.基于监控视频图像的车辆测速[J].中国图像图形学报,2005,10(2):192-196.
[39]Benjamin Coifman,David Beymer,Philip McLauchian,etal.A real-time computer vision system for vehicle tracking and traffic surveillance.Transportation Research.1998,PartC 6:271-288.
[40]Beymer D.,M.Philip,Ccoifman B.,etal.A real-time computer vision system for measuring traffic parameters.Proc.IEEE conf.computer vision and pattern recognition.1997,(7):496-501.
[41]Gupte S.,Masoud O.,etal.Detecting and classification of vehicles.IEEE transactions on intelligent transportation systems.2002,3(1):37-47.
[42]Anagnostopoulos.C.,Alexandropoulos.T..Intelligent traffic management through MPEG-7 vehicle flow surveillance.Proceedings-IEEE John Vincent Atanasoff 2006 International Symposium on Modern Computing,2006:202-207.
[43]Jun-Wei Hsieh,Shih-Hao Yu.Automatic traffic surveillance system for vehicle tracking and classification.IEEE Transactions on Intelligent Transportation Systems,2006:175-87.
[45]关向荣,任金昌.视频监控中背景的提取与更新算法[J].微电子学与计算机,2005,22(1):97-99
[46]艾海舟,乐秀宁.面向视觉监控实时跟踪的动态背景更新方法[J].计算机图形学报,2001,(9):104-106
[47]刘永信,魏平.复杂背景图像中检测动目标的一种方法[J].计算机工程与应用,2002,(23):22-24
[48]胡少鹏,李峰等.视频交通检测系统中车道背景更新算法研究[J],中国交通信息产业,2006,(7):108-111
[2]Tseng B.L.,Lin C.Y.,Smith J.R..Real-time video surveillance for traffic monitoring using virtual line analysis.2002 IEEE international coference,2002,2:541-544.
[3]Liu Y.,Payeur P.,Vision-based detection of activity for traffic control.IEEE CCECE 2003,Canadian,2003,2:1347-1350.
[4]Cucchiara R.,Piccardi M.,Prati A..Detecting moving objects,ghosts,and shadows in video streams.IEEE transactions on pattern analysis and machine intelligence.2003,(25):1337-1342.
[5]Yu M.,Jiang CRY.,Yu B.K..An integrative method for video based traffic parameter extraction in ITS.The 2000 IEEE Asia-Pacific Conference,2000:136-139.
[6]Wang J.M.,Chung Y.C.,Chang C.L.etal.Shadow detection and removal for traffic images.Proceeding of the 2004 IEEE international conference on Networking,sensing&control,Taipei,Taiwan,2004:649-654.
[7]杨佩昆.智能交通[M].上海:同济大学出版社,2002.
[8]《中国智能运输系统体系框架》专题组.中国智能运输系统体系框架[M].北京:人民邮电出版社,2003.
[9]史新宏,蔡伯根,穆建成.智能交通系统的发展[J].北方交通大学学报,2002,26(1):29-34.
[10]Rafael C.GONZALEZ,Richard E.WOODS.数字图像处理(第二版)[M].北京:电子工业出版社,2003.
[11]刘智勇.智能交通控制理论及其应用[M].北京:科学出版社,2003.
[12]阮秋琦等译.数字图像处理[M].北京:电子工业出版社,2005.
[13]章毓晋.图像处理和分析[M].北京:清华大学出版社,2000.
[14]向世明.Visual C++数字图像与图形处理[M].北京:电子工业出版社,2003.
[15]刘玮玮.Visual C++视频/音频开发实用工程案例精选[M].北京:人民邮电出版社,2005.
[16]徐慧等.Visual C++数字图像实用工程案例精选[M].北京:人民邮电出版社,2005.
[17]杨枝灵,王开.Visual C++数字图像获取、处理及实践应用[M].北京:人民邮电出版社,2003.
[18]胡小锋,赵辉.Visual C++/MATLAB图像处理与识别应用案例精选[M].北京:人民邮电出版社,2003.
[19]陈纯.计算机图像处理技术与算法[M].北京:清华大学出版社,2003.
[20]马骏.高速公路行车安全距离的分析与研究[J].西安公路交通大学学报,1998,18(4):90-94.
[21]Zhigang zhu,Guangyou Xu,Bo Yang,etal.A real-time vision system for automatic traffic monitoring.Image and Vision Computing,2000,18(20):781-794.
[22]张玲,叶海炳,何伟.一种基于边缘信息的改进车辆检测方法[J].重庆大学学报(自然科学版),2001,27(11):56-58.
[23]郁梅,蒋刚毅等.基于路面标记的车辆检测和计数[J].仪器仪表学报,2002,23(4):386-390.
[24]魏武,张起森等.基于计算机视觉和图像处理的交通流参数检测[J].信息与控制,2004,30(3):257-261.
[25]武想益,郑依展等.视频交通流检测演示系统[J].计算机工程,2003,29(21):151-153.
[26]王夏黎,周明全等.交通流视频检测系统的设计与实现[J].西北大学学报(自然科学版),2004,26(12):207-210.
[27]彭仁明,贺春林等.基于视频的车流量检测(自然科学版)[J].西华师范大学学报,2004,25(4):404-408.
[28]巨永锋,朱辉等.基于计算机视觉的车流量检测算法[J].长安大学学报(自然科学版),2004,24(1):92-95.
[29]王栓等.基于差分图像的多运动目标的检测与跟踪[J].中国图像图形学报,1999,4(6):470-475.
[30]张玲,陈丽敏等.基于视频的改进帧差法在车流量检测中的应用[J].重庆大学学报(自然科学版),2004,27(5):31-33.
[31]吴天昊,于前洋.用傅里叶相移特性估计位移[J].光电工程,2005,32(8):73-76.
[32]王东峰,邹谋炎.傅氏变换的自配准性质及其在纹理识别和图像分割中的应用[J].中国图象图形学报,2003,8(2):140-146.
[33]尹朝征.基于视频虚拟线圈的交通流参数检测[D].清华大学硕士学位论文.2002.
[34]李绪龙.基于图像处理技术的视频交通流信息采集系统研究[D].长安大学硕士学位论文.2003.
[35]王强.基于计算机视觉的交通流参数采集理论与方法研究[D].吉林大学硕士学位论文.2001.
[36]杜楠.复杂气候条件下的视频交通流量检测算法[D].哈尔滨工业大学硕士学位论文.2004.
[37]夏文龙.基于计算机视觉的汽车流量检测研究[D].四川大学硕士学位论文.2003.
[38]童剑军,邹明福.基于监控视频图像的车辆测速[J].中国图像图形学报,2005,10(2):192-196.
[39]Benjamin Coifman,David Beymer,Philip McLauchian,etal.A real-time computer vision system for vehicle tracking and traffic surveillance.Transportation Research.1998,PartC 6:271-288.
[40]Beymer D.,M.Philip,Ccoifman B.,etal.A real-time computer vision system for measuring traffic parameters.Proc.IEEE conf.computer vision and pattern recognition.1997,(7):496-501.
[41]Gupte S.,Masoud O.,etal.Detecting and classification of vehicles.IEEE transactions on intelligent transportation systems.2002,3(1):37-47.
[42]Anagnostopoulos.C.,Alexandropoulos.T..Intelligent traffic management through MPEG-7 vehicle flow surveillance.Proceedings-IEEE John Vincent Atanasoff 2006 International Symposium on Modern Computing,2006:202-207.
[43]Jun-Wei Hsieh,Shih-Hao Yu.Automatic traffic surveillance system for vehicle tracking and classification.IEEE Transactions on Intelligent Transportation Systems,2006:175-87.
[45]关向荣,任金昌.视频监控中背景的提取与更新算法[J].微电子学与计算机,2005,22(1):97-99
[46]艾海舟,乐秀宁.面向视觉监控实时跟踪的动态背景更新方法[J].计算机图形学报,2001,(9):104-106
[47]刘永信,魏平.复杂背景图像中检测动目标的一种方法[J].计算机工程与应用,2002,(23):22-24
[48]胡少鹏,李峰等.视频交通检测系统中车道背景更新算法研究[J],中国交通信息产业,2006,(7):108-111