摘要: 针对ViBe运动目标检测算法在实际环境中存在无法消除鬼影、阴影等干扰的问题,结合三帧差分、边缘检测等技术,提出了一种ViBe改进算法。预处理阶段通过三帧差分获得真实背景并消除鬼影,运动目标检测阶段结合先验知识和边缘检测方法获得真实的运动目标以消除阴影,目标描述与跟踪阶段运用像素标记分割方法得到目标描述并实现目标跟踪。实验结果表明,新方法在消除鬼影、阴影等干扰方面表现出了优越的性能,在交通监控实时视频流中具有理想的车辆检测和跟踪效果。
[1] Wixson L.Detecting Sailient Motion by Accumulating Directiona-lly Consistent flow[J].IEEE Trans.Pattern Analysis and Machine Intelligence,2000,22(8):774-780 [2] 张磊,史金飞,罗翔.运动目标检测的差分图像法分析研究[J].工业仪表与自动化装置,2007,5(3):7-12 [3] Barnich O,van Droogenbroeck M.ViBe:A powerful randomtechnique to estimate the background in video sequences[C]∥Proc.Int.Conf.Acoust,Speech Signal Process.2009:945-948 [4] Caseiro R,Henriques J,Martins P.Anonparametric riemannian framework on tensor field with application to foreground segmentation[C]∥ICCV 2011.2011:1-8 [5] McIvor A.Background subtraction techniques[C]∥Proc.ImageVis.Comput,Auckland.New Zealand,Nov.2000 [6] Benezeth Y,Jodoin P,Emile B.Review and evaluation of commonly-implemented background sub-traction algorithms[C]∥Proc.IEEEInt.Conf.Pattern Recognit.2008:1-4 [7] Parks D,Fels S.Evaluation of background subtraction algo-rithms with post-processing[C]∥Proc.IEEEInt.Conf.Adv.VideoSignal Based Surveillance.Santa Fe,New Mexico,2008:192-199 [8] Joshi K A,Thakore D G.A Survey on Moving Object Detection and Tracking in Video Surveillance System[J].International Journal of Soft Computing and Engineering,2012,2(3):44-48 [9] Pander R P,Mishra N D,Gulhane S.Detection of moving object with the help of motion detection alarm system in video survelliance[J].Journal of Signal and Image Processing,2012,2(3):118-121 [10] Barnich O,Van Droogenbroeck M.ViBe:A universal back-ground subtraction algorithm for video sequences[J].IEEE Transactions on Image Processing,2011,20(6):1709-1724 [11] Sonka M,Hlavac V,Boyle R.图像处理、分析与机器视觉[M].艾海舟,武勃,等译.北京:人民邮电出版社,2003 [12] 杨益军,赵春椿,汪文秉.一种适合于目标检测的图像分割方法[J].信号处理,2002,8(2):183-185 [13] 张新宇,刘广智,立建勤.一种图像分割的目标描述方法及实现[J].系统工程与电子技术,2003,25(2):219-222 [14] 陈忠碧,张启衡.一种适合于多目标检测的图像分割方法[J].光电工程,2004,1(5):34-37 [15] http://cvrr.ucsd.edu/aton/shadow/ |
No related articles found! |
|