基于鲁棒M估计和Mean Shift聚类的动态背景下运动目标检测

针对动态背景下运动目标的检测问题，提出了一种基于鲁棒M估计和Mean Shift聚类的目标检测新方法。首先，在考虑全局光照变化的情况下，构建鲁棒M估计器估计全局运动，以实现最小化相邻2帧图像中所有像素亮度的绝对残差和，根据M估计得到像素点权值，提取出代表局部运动信息的离群点；在离群点中均匀抽取网格点，然后利用Mean Shift聚类算法实现不同运动点的分割；根据聚类的结果生成凸包，准确分割出运动目标区域。实验结果表明，该方法能检测出动态背景下的多个运动目标区域，多目标检测准确度到达95%以上，并且只需两帧图像就可以准确检测并锁定运动目标，满足实时处理的要求，具有一定的工程意义.

Abstract

Focusing on the problem of moving objects detection in dynamic scenes, a novel algorithm based on robust M-estimator and mean shift clustering was proposed. First, considering the case of global illumination change, M-estimator was constructed to estimate the global motion in order to minimize the absolute residuals of pixels luminance between two adjacent frames. The structured outliers could be extracted according to the weight map of every pixel. Then the grid points were selected evenly from outliers and different point belong different moving object was clustered by mean shift algorithm. The convex hulls were generated under clustering results, to accurately segment the moving object regions. Experimental results show that this method can accurately detect multiple moving objects in dynamic scenes, and MODA can reach 95%. Besides, only two frames are needed to detect and lock the moving objects by this algorithm, which can meet real-time processing requirements and has a certain degree of engineering significance.

参考文献

[1] 周许超,屠大维,陈勇,等. 基于相位相关和差分相乘的动态背景下运动目标检测［J］. 仪器仪表学报,2010,31(5):980-983.

ZHOU Xu-chao,TU Da-wei,CHEN Yong,et al. Moving object detection under dynamic background based on phase-correlation and differential multiplication［J］. Chinese Journal of Scientific Instrument,2010,31(5):980-983.

[2] WU Q Z,JENG B S. Background subtraction based on logarithmic intensities［J］.Pattern Recognit Letter,2002,23(13):1529.

[3] JI X P,WEI Z Q,FENG Y W. Effective vehicle detection technique for traffic surveillance systems［J］.Journal of Visual Commun Image Represent,2006,17(3):647.

[4] 陈忠碧,张启衡,彭先蓉,等. 基于块估计的运动目标检测方法［J］. 光电工程,2006,33(6):15-19.

CHEN Zhong-bi,ZHANG Qi-heng,PENG Xian-rong ,et al.Detecting method for moving targets based on block estimation［J］. Opto-Electronic Engineering,2006,33(6):15-19.

[5] 汲清波,张兴周,刘云鹤. 动态背景的序列图像中运动小目标检测算法［J］. 计算机工程与应用,2007,43(35):93-95.

JI Qing-bo,ZHANG Xing-zhou,LIU Yun-he. Detection method for moving small target in infrared image sequences with dynamic background［J］. Computer engineering and applications,2007,43(35):93-95.

[6] 李哲,苏秀琴,杨小君,等. 红外运动小目标的检测［J］. 光子学报,2006,35(6):924-927.

LI Zhe,SU Xiu-qin,YANG Xiao-jun , et al. Small moving infrared target detection ［J］. Acta Photonica Sinica,2006,35(6):924-927.

[7] 齐美彬,汪巍,蒋建国,等. 动态场景下的快速目标检测算法［J］. 电子测量与仪器学报,2011,25(9):756-761.

QI Mei-shan,WANG Wei,JIANG Jian-guo,et al.Rapid moving object detection under a dynamic scene［J］. Journal of Electronic Measurement and Instrument,2011,25(9):756-761.

[8] HORNB K P，SCHUNCKB G. Determining optical flow［J］.ArtificalIntelligence,1981，17(1-3):185-203.

[9] 秦剑,陈钱,钱惟贤. 基于光流估计和自适应背景抑制的弱小目标检测［J］. 光子学报,2011,40(3):476-482.

QIN Jian,CHEN Qian,QIAN Wei-xian. A detection algorithm for dim and small infrared target based on the optical flow estimation and the adaptive background suppression［J］.Acta Photonica Sinica,2011,40(3):476-482.

[10] 罗寰,王芳,陈中起,等. 基于对称差分和光流估计的红外弱小目标检测［J］. 光学学报,2010,30(6):1715-1720.

LUO Huan,WANG Fang,CHEN Zhong-qi,et al.Infrared target detecting based on symmetrical displaced frame difference and optical flow estimation［J］.Acta Optica Sinica,2010,30(6):1715-1720.

[11] 李敏,范新南,张学武,等. 基于视觉认知机理的复杂动态背景下目标提取［J］. 光电子. 激光,2012,23( 2):366-373.

LI Ming,FAN Xin-nan,ZHANG Xue-wu,et al. A target extraction method in disordered and dynamic back-ground based on visual cognition theory［J］.Journal of Optoelectronics.Laser,2012,23( 2):366-373.

[12] LIU C,YUEN P C,QIU G. Object motion detection using information theoretic spatio-temporal saliency［J］. Pattern Recognition,2009,42(11):2897-2906.

[13] CHENG W H,CHU Wei-ta,WU Ja-ling. A visual attention based region-of-interest determination framework for video sequences［J］. IEICE transactions on information and systems,2005,88(7):1578-1586.

[14] 黄官远,严晖. 基于鲁棒M-估计器的全局运动估计方法［J］. 计算机工程,2009,35(3):235-236.

HUANG Guan-yuan,YAN Hui. Method for global motion estimation based on robust M-estimator［J］. Computer Engineering,2009,35(3):235-236.

[15] 程咏梅,王进行,魏坤,等. 基于加权 mean-shift 可见光/红外双通道目标跟踪［J］. 光子学报,2010,39(2):357-363.

CHEN Yong-mei,WANG Jin-xing ,WEI Kun ,et al. Visible/infrared dual-channel target tracking based on weighted Mean-shift ［J］. Acta Optica Sinica,2010,39(2):357-363.

[16] 吴文周,李利番,王结臣. 平面点集凸包 Graham 算法的改进［J］. 测绘科学,2010,35(6):123-125.

WU Wen-zhou,LI Li-fan,WANG Jie-chen. Graham planar convex hull algorithm［J］. Science of Surveying and Mapping,2010,35(6):123-125.

[17] KASTURI R,GOLDGOF D,SOUNDARARAJAN P,et al. Framework for performance evaluation of face,text,and vehicle detection and tracking in video:data,metrics,and protocol［J］. Pattern Analysis and Machine Intelligence,2009,31(2):319-336.

徐诚, 黄大庆. 基于鲁棒M估计和Mean Shift聚类的动态背景下运动目标检测[J]. 光子学报, 2014, 43(1): 0110001. XU Cheng, HUANG Da-qing. Detection of Moving Objects in Dynamic Scenes Based on Robust M-estimator and Mean Shift Clustering[J]. ACTA PHOTONICA SINICA, 2014, 43(1): 0110001.

基于鲁棒M估计和Mean Shift聚类的动态背景下运动目标检测

关于本站 Cookie 的使用提示

全站搜索

基于鲁棒M估计和Mean Shift聚类的动态背景下运动目标检测

相关论文

相关资讯

关于本站 Cookie 的使用提示

全站搜索