摘要
针对TLD(Tracking-Learning-Detection)算法在目标跟踪过程中容易出现跟踪失败的问题,在TLD算法的基础上引入了基于图像视觉伺服IBVS(Image Based Visual Servo)的理念,将图像视觉伺服控制应用到TLD算法中去,以此提高目标跟踪的鲁棒性,实现对目标的长时间跟踪,并且能够解决目标遮挡后的再次跟踪问题。实验表明,基于图像的视觉伺服能够有效控制目标跟踪的误差,将误差实时反馈给TLD算法,能够保证TLD算法跟踪的实时性与鲁棒性。
In order to solve the tracking failure problem of TLD( Tracking-Learning-Detection) algorithm in the process of target tracking,a IBVS( Image Based Visual Servo) concept was introduced to TLD algorithm,which makes the Image Visual Servo control combine with TLD algorithm to improve the robustness of target tracking. IBVS controls the error of target tracking and feeds the error back to TLD algorithm,which guarantees the real-time and robustness of TLD algorithm tracking and is able to solve the problem of tracking again after the object shelter. The experiments show that this method can be used for long time tracking,and the stability and efficiency are improved greatly.
引文
[1]高帆,吴国平,刑晨. TLD目标跟踪算法研究[J].电视技术,2013,37(11):70-74.
[2]MARFIL R,MOLINA T L,ODRIGUEZL J A. Real time object tracking using bounded irregular pyramids[J]. Pattern Recognition Letters,2007,28(9):985-1001
[3]夏瑜.视觉跟踪新方法及其应用研究[D].无锡:江南大学,2013.
[4]MALANOWSKI M. Algorithm for target tracking using passive radar[J]. International Journal of Electronics and Telecommunications,2012,58(4):345-350.
[5]龚小彪.基于TLD框架的目标跟踪算法研究[D].成都:西南交通大学,2014.
[6]吴忠文,费树岷.基于TLD目标跟踪算法的优化[J].工业控制计算机,2018,31(4):77-79.
[7]谷文华,蒋念平,李振兴.基于粒子滤波的TLD目标跟踪算法[J].电子科技,2015,28(12):45-47,51.
[8]KALAL Z,MALANOWSKI K,MATAS J. Tracking-Learning-Detection[J]. IEEE Trans on Pattern Analysis&Machine Intelligence,2012,34(7):1409-1422.
[9]KALAL Z,MATAS J,MALANOWSKI K. P-N Learning:bootstrapping binary classifiers by structural constrains[J]. IEEE Computer Society Conference On Computer Vision and Pattern Recognition,2010,359(4):49-56.
[10]康振波.基于图像的机器人双目视觉伺服系统控制研究[D].秦皇岛:燕山大学,2012.
[11]贾丙西,刘山,张凯祥,等.机器人视觉伺服研究进展:视觉系统与控制策略[J].自动化学报,2015,41(5):861-873.
