摘要
基于视频的交通事件检测系统能对道路上出现的车辆逆行、车辆抛锚、抛落物等异常事件进行实时检测,对避免二次事故的发生起了重要的作用,是目前国内外研究的一个热点,有着广阔的应用前景。
本文对交通事件检测算法进行了研究,主要内容包括:
(1)背景估计。在分析了传统的背景估计方法的基础上提出一种新的背景估计方法。该方法把图像分成大小一定的块,通过统计在时间序列上图像的每个块内的方差与均值落在由均值与方差组成的二维平面上的各个区域的概率来对背景图像进行分块估计。
(2)图像的二值化分割。在分析了传统二值化分割法的基础上提出了一种新的二值化分割方法。该方法把图像分成大小一定的块,通过对前景图像与背景图像的每个块内的像素进行基于最小二乘法的灰度拉升,改善了块内像素灰度分布的均衡性,然后通过前景图像每个块与背景图像中相应块中像素灰度差的绝对值之和的运算,进行以块为单位的二值化分割。
(3)事件检测。在介绍了传统的事件检测方法的基础上提出了一种新的事件检测方法。该方法基于有目标的块在时间序列上的累积值产生的分布规律来进行异常事件的检测。
本文所设计出的算法在不同交通流量的场景中进行了实验,实验结果表明,该算法能够很好的解决交通流量较大的情况下常规背景估计法不能解决的在场景远处误估计率高的问题,并能有效的克服影子、车周边灰度值较低的区域对二值化分割的影响,新提出的事件检测算法在运算量上满足了实时检测的要求。初步的应用表明:该算法具有良好的适用性与应用前景。
Video-based traffic incident detection system can detect the broken down vehicles,retrograde vehicles,thrown objects and other abnormal incidents on the road real-time,which plays an important role in avoiding secondary accidents,it's a research focus at home and abroad currently and have broad application prospects.
This paper studies the traffic incident detection lgorithms,the main contents are as follows:
(1) Background Estimation. On the basic of analyzing the traditional background estimation methods,a new background estimation method is proposed.This method divide a frame of image into a number of blocks with fixed size. Calculating the mean and the variance of each block in the time series images,by Statistics the probability of the mean and the variance fall into each area in the two-dimensional palne of which consists the mean and the variance ,the background image is estimated in blocks.
(2) Image Binarization Segmentation. On the basic of analyzing the traditional image binarization methods,a new image binarization method is proposed.This method divide a frame of image into a number of blocks with fixed size.By changing every pixel gray value in each block in the froground and background images base on the least square method,improve the distribution of the pixel gray balance in a block,and then subtract the pixel in a every block in the forground images from the corresponding pixel in the background images,calculating the sum of the absolute value of the difference in each block,and the image binarization segmentation completed in blocks.
(3) Incident Detection. On the basic of analyzing the traditional incident detection methods,a new Incident Detection method is proposed.This method detect the abnormal events bases on cumulative value distribution Of the targeted blocks in the time series.
The designed algorithm in this paper experimented in different traffic flow scenarios shows that the new proposed incident algorithm meet the real time detection requirements. The initial application shows that the algorithm has applicability and with good prospects.
引文
[1]Shuming Tang and Haijun Gao,Traffic-Incident Detection Algorithm Based on Nonparametric Regression.IEEE TRANSACTIONS.ON INTELGENT TRANSPORTATION SYSTEMS.VOL.6
[2]Chin-Der Wanna,Ming HuiLin.Data Fusion Methods for Accuracy Improvement in Wireless Location Systems.WCNC 2004/IEEE Communications Society.2004
[3]Michael D.Fontaine,and Brian L.Smith.Impact of system parameters on the accuracy of wireless loacation technology-based traffic monitoring system.2004 IEEE Intelligent Transportation Systems Conference Washington,D.C,USA.October 34,2004
[4]Carol.H.Walters,Poonam.B.Wiles,Scott.A.Cooner.Incident Detection Primarily By Cellular Phones-An Evaluation of a System for Dallas,Texas[J].78 th Annual Meeting of the Transportation Research Borad,1999.
[5]向怀坤,柴旭东,廖明军等.给予 GPS,GIS 与 GSM 开发车辆监控调度系统的研究[J].汽车工程,2002,34(3):195-199.
[6]杨飞,裘炜毅.给予手机定位的实时交通数据采集技术[J].城市交通,2005,13(4):63-68
[7]南振岐.数字城市公众地理信息系统之智能交通车辆 GPS 监控调度系统方案[J].中国科技信息,2005,(19):35239.
[8]李传志,胡如夫.公路交通事件检测技术的研究与展望[J].交通科技与经济,2008,(5):68-71.
[9]汪勤,黄山,张洪斌,杨权,张建军.基于视频图像处理的交通事件检测系统[J].计算机应用2008,(7):1887-1889.
[10]王宏义.基于视频图像的交通事件检测系统研究与开发[J].公路与汽运,2007,5(3):57-60.
[11]汤淑明,王坤峰,李元涛.基于视频的交通事件自动检测技术综述[J]公路交通科技,2006,(8):116-121.
[12]郭永涛,宋焕生.视频交通监控系统中背景提取算法[J].视频技术应用与工程,2006,12(1):91-93.
[13]刘飞,张鹏林.基于卡尔曼滤波的时间序列影像背景估计[J].地理空间信息,2005,6(3):12-13.
[14]潘建寿,王琳.一种基于分级分块处理的背景估计算法[J].光子学报,2005,34(12):1898-1900.
[15]彭强,李华.基于块直方图分析的视频背景提取方法[J].西南交通大学学报,2006,41(1):48-53.
[16]黄浩军,胡鹏.交通视频检测系统中背景提取的优化算法[J].湖北汽车工业学院学报,2006,20(4):52-55.
[17]Otsu N.A threshold selection method from gray—levelhistograms[J].IEEE Trans Systems,Man and Cy—bemetics,1979,9(1):62-66.
[18]BelYisen J.Dynamic thresholding of gray-level images[A].Proc.8th Int.Corr.on Pattem Recognition[C].Los Angeles:IEEE Computer Society Press,1986:1251-1255.
[19]胡建华,扬帆.基于分块差分与背景减除的运动检测[J].电子测量技术,2007,30(10):48-53.
[20]陈浩,陈泉林.智能交通监视系统中运动目标提取算法研究[J].计算机应用与软件,21(4):74-75.
[21]朱瑾瑜,刘海燕.基于连通性检测的视频监控运动目标提取[J].视频技术应用与工程,31(10):78-80.
[22]肖志坚.H.264 帧间编码技术的分析与改进[D].陕西省西安市:西安电子科技大学.2005.
[23]彭小宁,杨明.基于局部图金字塔的不规则块匹配视频分割方法[J].计算机科学.,2008
[24]Dorin Comaniciu.Kernel-Based Object Tracking[J].IEEE transactions on pattern analysis and machine intelligence,v 25,n 5.2003
[25]Visvanathan Ramesh.Real-Time Tracking of Non-Rigid Objects using Mean Shift[J].IEEE computer vision and pattern recognition.2000
[26]G.R.Bradski.Computer vision face tracking as a component of a perceptual user interface[J].IEEE Computer Vision.1998,p 214-219
[27]Y.Bar-Shalom,T.Fortmann.Tracking and Data Association[J].Academic Press.1988
[28]S.Birchfield.Elliptical head tracking using intensity gradients and color histograms[J].IEEE Computer Vision and Pattern Recognition.1998,p 232-237
[29]N.J.Gordon,D.J.Salmond,A.F.M.Smith.Novel approach to nonlinear/non-Gaussian Bayesian state estimation[J].IEEE proceedings.1993
[30]Micheal Isard,Andrew Blake.Condensation-conditinal density propagation for visual tracking[J].Computer Vision.1998
[31]Bascle,B,Deriche,R.Region tracking through image sequences[C].Proc.5th Int.Conf.on Computer Vision 1995.p 302-307
[32]Black,M.,Yacoob,Y.Tracking and recognizing rigid and non-rigid facial motions using local parametric models of image motion[C].Proc.5th Int.Conf.on Computer Vision.1995.p 374-381
[33]郭博学.运动车辆视频分割与跟踪技术研究[D].陕西省西安市 长安大学,2009
[34]Surendra Gupte,Osama Masoud,Robert F.K.Martin,Nikolaos P.Papanikolopoulos.Detection and Classification of Vehicles[J].IEEE transactions on intelligent transportation systems,v 3,n 1.2002