公路交通事件自动检测算法研究
|
摘要
为了及时检测出道路上发生的各类交通事件,降低由于交通事件所带来的人员伤亡和财产损失,避免类似事件的再次发生,为各职能部门执法提供科学依据,所以,建立交通事件检测系统是非常必要的。而交通事件检测算法又是交通事件检测系统的核心部分,它的发展至关重要。
本文先介绍了国内外交通事件自动检测算法的研究现状,然后对经典事件检测算法进行了概述,并通过算法评价指标对其进行了性能评价。在此基础上,综合考虑了我国及四川省公路的现状,提出了一种适于我国公路的基于图像检测的交通事件检测算法。该算法设计思想是以图像预处理为基础,以运动车辆的检测和跟踪算法为重点。运动目标的检测方面,提出了一种新的思路,即把对称差分法和背景差分法结合起来共同检测场景内的运动目标。运动车辆跟踪方面,是将卡尔曼滤波和模板匹配法结合在一起实现目标的跟踪。本论文的另一个研究方面是智能检测系统软件的设计,通过对监控图像的分析处理,实现多车辆轨迹跟踪、车辆禁行、禁停区域违规,车辆逆行违规等多种用户可以自主设定的功能。
实验方面,本文对图像预处理的四个方面进行了Matlab仿真,对提出的新的检测和跟踪算法列出了实验结果,并对结果作出了分析说明。本文中涉及到的研究思路可以说是交通事件自动检测方法的一种探索,对今后的交通事件自动检测具有一定的参考意义。
In order to detect and deal with all kinds of traffic events on roads so as to decrease the death rate and the loss of wealth, to prevent similar incidents to happen again. Therefore, establishing the Traffic Events Detection System(TEDS) is very necessary. As the core of the TEDS, algorithm of traffic events detection is worth studying.
This thesis first introduces the current research situation of algorithms for automatic detection of traffic events at home and abroad, and then summarizes the classic events detection algorithm, and through the algorithm evaluation index to evaluate classic events detection algorithm. On this basis, considering the status of highway in China and Sichuan Province, I proposes a suitable algorithm based on image for China's national highway traffic events detection. The algorithm based on Image Pre-processing, focusing on the detection and tracking algorithm of the moving vehicle. Moving target detection, this paper proposes a new train of thought, namely combine symmetric difference method and background subtraction method detecting moving objects within the scene together. Motor vehicle tracking, kalman filter and template matching method is combined to realize target tracking Another major aspect of this paper is intelligent detection system. By monitoring the image analysis and processing to achieve the traffic violations forbidden line region,the vehicle retrograde violations and so on. Users can set their own function
Finally, on the basis of the objectives of the system, the paper uses Matlab technology to emulate. The research ideas and the algorithm proposed in this paper is a method exploration of traffic automatic incident detection algorithm and may provide certain theoretical reference.
本文先介绍了国内外交通事件自动检测算法的研究现状,然后对经典事件检测算法进行了概述,并通过算法评价指标对其进行了性能评价。在此基础上,综合考虑了我国及四川省公路的现状,提出了一种适于我国公路的基于图像检测的交通事件检测算法。该算法设计思想是以图像预处理为基础,以运动车辆的检测和跟踪算法为重点。运动目标的检测方面,提出了一种新的思路,即把对称差分法和背景差分法结合起来共同检测场景内的运动目标。运动车辆跟踪方面,是将卡尔曼滤波和模板匹配法结合在一起实现目标的跟踪。本论文的另一个研究方面是智能检测系统软件的设计,通过对监控图像的分析处理,实现多车辆轨迹跟踪、车辆禁行、禁停区域违规,车辆逆行违规等多种用户可以自主设定的功能。
实验方面,本文对图像预处理的四个方面进行了Matlab仿真,对提出的新的检测和跟踪算法列出了实验结果,并对结果作出了分析说明。本文中涉及到的研究思路可以说是交通事件自动检测方法的一种探索,对今后的交通事件自动检测具有一定的参考意义。
In order to detect and deal with all kinds of traffic events on roads so as to decrease the death rate and the loss of wealth, to prevent similar incidents to happen again. Therefore, establishing the Traffic Events Detection System(TEDS) is very necessary. As the core of the TEDS, algorithm of traffic events detection is worth studying.
This thesis first introduces the current research situation of algorithms for automatic detection of traffic events at home and abroad, and then summarizes the classic events detection algorithm, and through the algorithm evaluation index to evaluate classic events detection algorithm. On this basis, considering the status of highway in China and Sichuan Province, I proposes a suitable algorithm based on image for China's national highway traffic events detection. The algorithm based on Image Pre-processing, focusing on the detection and tracking algorithm of the moving vehicle. Moving target detection, this paper proposes a new train of thought, namely combine symmetric difference method and background subtraction method detecting moving objects within the scene together. Motor vehicle tracking, kalman filter and template matching method is combined to realize target tracking Another major aspect of this paper is intelligent detection system. By monitoring the image analysis and processing to achieve the traffic violations forbidden line region,the vehicle retrograde violations and so on. Users can set their own function
Finally, on the basis of the objectives of the system, the paper uses Matlab technology to emulate. The research ideas and the algorithm proposed in this paper is a method exploration of traffic automatic incident detection algorithm and may provide certain theoretical reference.
引文
[1]中建标公路委员会编.公路工程技术标准.人民交通出版社.2004.
[2]小溪.四川省公路发展概况.中国网china.com.cn.2009.
[3]李丹.多方会商破解“顽疾”保畅通.四川日报.2008.
[4]徐学才.高速公路事件管理系统设计及算法实现.西南交通大学硕士论文.2004.
[5]张敬磊,王晓原.交通事件检测算法研究进展.武汉理工大学学报.2005.
[6]代磊磊,姜桂艳,韩国华.高速公路事件自动检测算法综述.ITS通讯.2004.
[7]符锌砂.广州市中心区道路交通紧急救援系统研究报告[R].2004：55.
[8]朱毅,张朝亮,周君.计算机与数字工程.计算机监控系统中的图像预处理技术.2007.
[9]冈萨雷斯.数字图像处理第二版中文版.电子工业出版社.2004.
[10]董梁；王忠民.图像边缘检测算法研究.现代电子技术.2007.
[11]段瑞玲,李庆祥,李玉和.图像边缘检测方法研究综述.光学技术.2005.
[12]史新宏,蔡伯根.高速公路自动事件检测算法[J].交通运输系统工程与信息,2001,1(4)：306-310.
[]3] 陈朝阳,张桂林.基于图像对称差分运算的运动小目标检测方法.华中理工大学学报.1998.
[14]Jain R, NagelHH. On the Analysis of Accumulative Difference of Picture from Image Sequences of Real World Scenes. IEEE Trans. PAMI,1979, PAMI 1 (1):206-214.
[15]Kapur J N. A New Method fo r Gray-Level Picture Threshold Using the Entropy of the Histogram. CVGIP,1985,29 (1):273-285.
[16]WikLand J, Gr and G. T racking of Multiple Moving Objects. in: Cappellini V ed. Time Varying Image Processing and Moving Objects Recognition. [s.1.]:Eisevier Science Publishers,1987.
[17]宋莹,陈科.基于图像分块的边缘检测算法.计算机工程.2010.7.
[18]徐以美,郭宝龙,张晋.基于像素分类的运动目标检测算法.计算机工程.2008,6.
[19]WANG Xiao-ping, ZHANG Li-jie, CHANG Ji. Improved method of moving objects detection based on single-gaussian background model. Computer Engineering and Applications,2009,45 (21):118-120.
[20]胡学刚,刘忠振.基于高斯背景模型的车辆检测改进算法.计算机工程与设计.2011,12.
[21]张敬磊,王晓原.交通事件检测算法研究进展[N].武汉理工大学学报(交通科学与工程版),2005,29(2)：215-218.
[22]张洋,叶冬升,朱岩.高速公路交通事件检测系统研究[J].交通与计算机,2003(3)：6-9.
[23]伍友龙.基于图像分析的高速公路交通事件检测算法研究[D].长沙：长沙理工大学,2005.
[24]王东升.视频运动目标检测及其实时处理系统研究[D].成都：电子科技大学,2004.
[25]乔光军,杨兆选.基于虚拟线的交通信息视频检测技术及应用[J].电子技术应用,2005,9：23-25.
[26]彭强,李华.基于块直方图分析的视频背景提取方法[N].西南交通大学学报,2006,41(1)：48-53.
[27]吴祖林,沈庆宏,都思丹等.背景提取基础上运动车辆视频检测[z].交通与计算机,2003,6(21)：18-20.
[28]FU Ping, FANG Shuai, XU Xinhe, XUE Dingyu. Approach for Shadow Removal of Moving Object Detection in Surveillance S ystem. Computer Engineering.2007,5.
[29]周欣,黄席越.基于卡尔曼滤波的车辆跟踪技术.四川师范大学学报(自然科学版).2008,11.
[30]WANG Dan-ling, LU Yong-quan, JIA Xiao-jie, ZHANG Qin. Real-time Particle Filter Tracking Algorithm and Implementation.School of Information and Engineering, Communication University of China Beijing 100024, China.2009,9.
[31]吴志刚.智能监控系统中目标检测算法的研究[D].武汉理工大学硕士论文.2009.
[32]M. Naylor, C. I. Annotated Digital Video for Intelligent Surveillance and Optimized Retrieval [J]. University of Reading and Kingston University:UK.2003.
[33]郑江滨,张艳宁,冯大淦等.视频监视中运动目标的检测与跟踪算法[J].系统工程与电子技术.2002,10：34-37.
[34]Aydin, Leven Onural, Michael Wollborn, Roland Mech. Image Sequence Analysis for Emerging Interactive Multimedia Services [J]. The European COST21 1 Framework, IEEE Trans. On CSVT. 1998.8(7):802-813.
[35]D. S. Zhang, G Lu. Segmentation of Moving Objects in Image Sequence: AReview [J]. Circuits Systems and Signal Processing (Special Issue on Multimedia Communication Services).2001,20(2):143-183.
[36]JunKi.IOm, Ho Suk Lee. Real-time Pre-processing and Video Object Segmentation for High Compression and Content-based MPEG-4 coding[C]. Processing of the Data Compression Conference (DCC 2003).2003,3:424-434.
[37]刘勃,周荷琴,魏铭旭.基于颜色和运动信息的夜间车辆检测方法[J].中国图象图形学报,2005,10(2)：187-191.
[38]Haddon, J. E, Boyce. LF. Image segmentation by unifying region and boundary information [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence.1990,12(10):929-948.
[39]侯伟.视频对象分割技术及应用[D].浙江工业大学硕士论文.2008.
[40]RAFAEL C.GONZALEZ RICHARD E.WOODS. Digital Image Processing (Second Edition) [M]电子工业出版社.2005：420-450.
[41]Gloyer B, Aghajan HK, Siu KY Kailath T. Video-Based freeway monitoring system using recursive vehicle tracking [J]. IEEE,1995.
[42]Sen-Ching S, Cheung, Chandrika Kamath. Robust techniques for background subtraction in urban traffic video [J]. San Jose:Video Communications and Image Processing, SPIE Electronic Imaging. 2004.
[43]Friedman N, Russell S. Image segmentation in Video Sequences:A Probabilistic Approach. Proc. Thirteenth Conf. on Uncertainty in Artificial Intelligence (UAI 97).
[44]Lee Dar-Shyang. Effective Gaussian Mixture Learning for Video Background Subtraction. IEEE Transactions on Pattern Analysis And Machine Intelligence.2005,27(5):34-36.
[45]翟永前,于惠芬,赵力,邹采荣.一种基于视频技术的车辆跟踪方法.电气电子教学学报.2011.
[46]刘鸿伟.数字化视频监控系统中运动目标检测关键技术及实现[D].北京科技大学硕士学位论文.2005.
[47]朱明旱,罗大庸,曹倩霞.帧间差分与背景差分相融合的运动目标检测算法 [2].计算机测量与控制,2005,13(3)：215-217.
[48]邹晓春.基于快速块匹配的图像序列运动估计技术研究[D].西安：西北工业大学,2005.
[49]刘波,李宪利.视觉信息的检测与跟踪技术的研究[J].计算机工程,2000(0)：86-89.
[50]王宁.基于视频的车辆跟踪与交通事件检测[D].成都：四川大学,2005.
[51]颜文才.视频动目标检测与跟踪算法研究[D].合肥：国防科学技术大学,2004.
[52]周志宇,汪亚明,黄文清.基于动态图像序列的运动目标跟踪[s].浙江工程学院学报,2002(9)：165-170.
[53]佟守愚,程三伟,李江,付萍.基于视频技术的交通违章处理系统的设计与实现.计算机测量与控制.2005.
[2]小溪.四川省公路发展概况.中国网china.com.cn.2009.
[3]李丹.多方会商破解“顽疾”保畅通.四川日报.2008.
[4]徐学才.高速公路事件管理系统设计及算法实现.西南交通大学硕士论文.2004.
[5]张敬磊,王晓原.交通事件检测算法研究进展.武汉理工大学学报.2005.
[6]代磊磊,姜桂艳,韩国华.高速公路事件自动检测算法综述.ITS通讯.2004.
[7]符锌砂.广州市中心区道路交通紧急救援系统研究报告[R].2004：55.
[8]朱毅,张朝亮,周君.计算机与数字工程.计算机监控系统中的图像预处理技术.2007.
[9]冈萨雷斯.数字图像处理第二版中文版.电子工业出版社.2004.
[10]董梁；王忠民.图像边缘检测算法研究.现代电子技术.2007.
[11]段瑞玲,李庆祥,李玉和.图像边缘检测方法研究综述.光学技术.2005.
[12]史新宏,蔡伯根.高速公路自动事件检测算法[J].交通运输系统工程与信息,2001,1(4)：306-310.
[]3] 陈朝阳,张桂林.基于图像对称差分运算的运动小目标检测方法.华中理工大学学报.1998.
[14]Jain R, NagelHH. On the Analysis of Accumulative Difference of Picture from Image Sequences of Real World Scenes. IEEE Trans. PAMI,1979, PAMI 1 (1):206-214.
[15]Kapur J N. A New Method fo r Gray-Level Picture Threshold Using the Entropy of the Histogram. CVGIP,1985,29 (1):273-285.
[16]WikLand J, Gr and G. T racking of Multiple Moving Objects. in: Cappellini V ed. Time Varying Image Processing and Moving Objects Recognition. [s.1.]:Eisevier Science Publishers,1987.
[17]宋莹,陈科.基于图像分块的边缘检测算法.计算机工程.2010.7.
[18]徐以美,郭宝龙,张晋.基于像素分类的运动目标检测算法.计算机工程.2008,6.
[19]WANG Xiao-ping, ZHANG Li-jie, CHANG Ji. Improved method of moving objects detection based on single-gaussian background model. Computer Engineering and Applications,2009,45 (21):118-120.
[20]胡学刚,刘忠振.基于高斯背景模型的车辆检测改进算法.计算机工程与设计.2011,12.
[21]张敬磊,王晓原.交通事件检测算法研究进展[N].武汉理工大学学报(交通科学与工程版),2005,29(2)：215-218.
[22]张洋,叶冬升,朱岩.高速公路交通事件检测系统研究[J].交通与计算机,2003(3)：6-9.
[23]伍友龙.基于图像分析的高速公路交通事件检测算法研究[D].长沙：长沙理工大学,2005.
[24]王东升.视频运动目标检测及其实时处理系统研究[D].成都：电子科技大学,2004.
[25]乔光军,杨兆选.基于虚拟线的交通信息视频检测技术及应用[J].电子技术应用,2005,9：23-25.
[26]彭强,李华.基于块直方图分析的视频背景提取方法[N].西南交通大学学报,2006,41(1)：48-53.
[27]吴祖林,沈庆宏,都思丹等.背景提取基础上运动车辆视频检测[z].交通与计算机,2003,6(21)：18-20.
[28]FU Ping, FANG Shuai, XU Xinhe, XUE Dingyu. Approach for Shadow Removal of Moving Object Detection in Surveillance S ystem. Computer Engineering.2007,5.
[29]周欣,黄席越.基于卡尔曼滤波的车辆跟踪技术.四川师范大学学报(自然科学版).2008,11.
[30]WANG Dan-ling, LU Yong-quan, JIA Xiao-jie, ZHANG Qin. Real-time Particle Filter Tracking Algorithm and Implementation.School of Information and Engineering, Communication University of China Beijing 100024, China.2009,9.
[31]吴志刚.智能监控系统中目标检测算法的研究[D].武汉理工大学硕士论文.2009.
[32]M. Naylor, C. I. Annotated Digital Video for Intelligent Surveillance and Optimized Retrieval [J]. University of Reading and Kingston University:UK.2003.
[33]郑江滨,张艳宁,冯大淦等.视频监视中运动目标的检测与跟踪算法[J].系统工程与电子技术.2002,10：34-37.
[34]Aydin, Leven Onural, Michael Wollborn, Roland Mech. Image Sequence Analysis for Emerging Interactive Multimedia Services [J]. The European COST21 1 Framework, IEEE Trans. On CSVT. 1998.8(7):802-813.
[35]D. S. Zhang, G Lu. Segmentation of Moving Objects in Image Sequence: AReview [J]. Circuits Systems and Signal Processing (Special Issue on Multimedia Communication Services).2001,20(2):143-183.
[36]JunKi.IOm, Ho Suk Lee. Real-time Pre-processing and Video Object Segmentation for High Compression and Content-based MPEG-4 coding[C]. Processing of the Data Compression Conference (DCC 2003).2003,3:424-434.
[37]刘勃,周荷琴,魏铭旭.基于颜色和运动信息的夜间车辆检测方法[J].中国图象图形学报,2005,10(2)：187-191.
[38]Haddon, J. E, Boyce. LF. Image segmentation by unifying region and boundary information [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence.1990,12(10):929-948.
[39]侯伟.视频对象分割技术及应用[D].浙江工业大学硕士论文.2008.
[40]RAFAEL C.GONZALEZ RICHARD E.WOODS. Digital Image Processing (Second Edition) [M]电子工业出版社.2005：420-450.
[41]Gloyer B, Aghajan HK, Siu KY Kailath T. Video-Based freeway monitoring system using recursive vehicle tracking [J]. IEEE,1995.
[42]Sen-Ching S, Cheung, Chandrika Kamath. Robust techniques for background subtraction in urban traffic video [J]. San Jose:Video Communications and Image Processing, SPIE Electronic Imaging. 2004.
[43]Friedman N, Russell S. Image segmentation in Video Sequences:A Probabilistic Approach. Proc. Thirteenth Conf. on Uncertainty in Artificial Intelligence (UAI 97).
[44]Lee Dar-Shyang. Effective Gaussian Mixture Learning for Video Background Subtraction. IEEE Transactions on Pattern Analysis And Machine Intelligence.2005,27(5):34-36.
[45]翟永前,于惠芬,赵力,邹采荣.一种基于视频技术的车辆跟踪方法.电气电子教学学报.2011.
[46]刘鸿伟.数字化视频监控系统中运动目标检测关键技术及实现[D].北京科技大学硕士学位论文.2005.
[47]朱明旱,罗大庸,曹倩霞.帧间差分与背景差分相融合的运动目标检测算法 [2].计算机测量与控制,2005,13(3)：215-217.
[48]邹晓春.基于快速块匹配的图像序列运动估计技术研究[D].西安：西北工业大学,2005.
[49]刘波,李宪利.视觉信息的检测与跟踪技术的研究[J].计算机工程,2000(0)：86-89.
[50]王宁.基于视频的车辆跟踪与交通事件检测[D].成都：四川大学,2005.
[51]颜文才.视频动目标检测与跟踪算法研究[D].合肥：国防科学技术大学,2004.
[52]周志宇,汪亚明,黄文清.基于动态图像序列的运动目标跟踪[s].浙江工程学院学报,2002(9)：165-170.
[53]佟守愚,程三伟,李江,付萍.基于视频技术的交通违章处理系统的设计与实现.计算机测量与控制.2005.