计算机科学 ›› 2016, Vol. 43 ›› Issue (5): 288-293.doi: 10.11896/j.issn.1002-137X.2016.05.055
范菁,阮体洪,吴佳敏,董天阳
FAN Jing, RUAN Ti-hong, WU Jia-min and DONG Tian-yang
摘要: 从高速交通监控视频中提取的车辆轨迹数据可以用于分析和识别车辆行为。由于从高速监控视频中提取的车辆轨迹中只有少量的变道、超车等类型轨迹,采用经典的最长公共子串(LCSS)相似度和谱聚类等算法无法有效地区分轨迹数据中所有类型的轨迹;此外,在车辆行为识别方面,常用的隐马尔科夫(HMM)轨迹模型忽略了负样本的影响,且仅用最大似然值进行分类,存在较高的误识别率。为了解决这些问题,分析和研究了高速监控视频中车辆轨迹数据的特点,提出了一种基于二次谱聚类和HMM-RF混合模型的车辆行为识别方法。该方法利用轨迹曲率来识别具有曲线轨迹特征的超车轨迹,利用倾角相似度和谱聚类算法来识别非曲线轨迹中的变道轨迹,并将得到的所有聚类簇用LCSS和谱聚类算法进行再聚类,从而有效地区分超车、变道以及直行轨迹等。在进行车辆行为识别时,该方法通过将不同HMM模型的多维概率输出作为随机森林RF模型的输入来识别多类型轨迹以替代最大似然值分类,提高了行为识别的正确率。为了验证方法的有效性,在不同数据集下进行实验,结果表明轨迹聚类的平均准确率为96%,而行为识别的平均准确率是89.3%,算法具有较高的准确率和鲁棒性。
[1] Atev S,Miller G,Papanikolopoulos N P.Clustering of vehicletrajectories [J].IEEE Transactions on Intelligent Transportation Systems,2010,11(3):647-657 [2] Morris B T,Trivedi M M.Understanding vehicular traffic behavior from video:a survey of unsupervised approaches[J].Journal of Electronic Imaging,2013,22(4):6931-6946 [3] Morris B T,Trivedi M M.Trajectory learning for activity understanding:Unsupervised,multilevel,and long-term adaptive approach[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2011,33(11):2287-2301 [4] Morris B,Trivedi M.Learning trajectory patterns by clustering:Experimental studies and comparative evaluation[C]∥IEEE Conference on Computer Vision and Pattern Recognition,2009(CVPR 2009).IEEE,2009:312-319 [5] Morris B,Trivedi M.An adaptive scene description for activity analysis in surveillance video[C]∥19th International Confe-rence on Pattern Recognition,2008(ICPR 2008).IEEE,2008:1-4 [6] Morris B T,Trivedi M M.Learning and classification of trajectories in dynamic scenes:A general framework for live video analysis[C]∥IEEE Fifth International Conference on Advanced Video and Signal Based Surveillance,2008(AVSS’08).IEEE,2008:154-161 [7] Hu W,Xiao X,Fu Z,et al.A system for learning statistical motion patterns[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2006,28(9):1450-1464 [8] Piciarelli C,Micheloni C,Foresti G L.Trajectory-based anomalous event detection[J].IEEE Transactions on Circuits and Systems for Video Technology,2008,18(11):1544-1554 [9] Breiman L.Random forests [J].Machine Learning,2001,45(1):5-32 [10] Zelnik-Manor L,Perona P.Self-tuning spectral clustering[C]∥Advances in Neural Information Processing Systems.2004:1601-1608 [11] Chen J,Wang R,Liu L,et al.Clustering of trajectories based onHausdorff distance[C]∥ 2011 International Conference on Electronics,Communications and Control(ICECC).IEEE,2011:1940-1944 [12] Hervieu A,Bouthemy P,Le Cadre J P.A HMM-based method for recognizing dynamic video contents from trajectories[C]∥IEEE International Conference on Image Processing,2007(ICIP 2007).IEEE,2007,4:533-536 [13] Wang Y,Yang C,Wu X,et al.Kinect based dynamic hand gesture recognition algorithm research[C]∥2012 4th International Conference on Intelligent Human-Machine Systems and Cybernetics(IHMSC).IEEE,2012:274-279 |
No related articles found! |
|