基于帧间差分和水平集的运动目标探测跟踪方法  

Moving Targets Detection and Tracking Based on Frame Difference and Level Set

作　　者：何曦[1] 瞿建荣[1] 卢晓燕[1] 刘培桢[1] 王娇颖[1] 

机构地区：[1]西安应用光学研究所,陕西西安710065

出　　处：《探测与控制学报》2015年第1期27-30,共4页Journal of Detection & Control

摘　　要：针对运动目标探测中,帧间差分法容易出现目标拉伸和空洞的问题,提出了基于帧间差分和水平集的运动目标探测跟踪方法。该方法首先利用帧差法获取运动目标的大概位置,再利用水平集得到目标的完整区域,最后利用直方图匹配的方式对目标进行关联,实现实时目标跟踪,这样既解决了帧间差分法中目标拉伸和空洞问题,又克服了水平集自身无法利用目标运动信息和初始化不准确时收敛速度慢的问题,提高了探测的准确性和跟踪的稳定性。实验结果表明,该方法能够在实时条件下,实现目标的精确探测和稳定跟踪。Stretching and hole of moving targets may occur during the targets detection by frame difference method.This paper proposed a detection and tracking method of moving targets based on the frame difference and level set.Firstly,frame difference was used to get the approximate location of a moving target;then,the complete area was obtained by using the level set;finally,histogram matching helped associating with the targets,achieving the real-time tracking.This solution not only solved stretching and hole of frame difference,but also overcame the problems that the level set itself could not use target motion information and slow convergence speed of initializing the inaccurate.The detection accuracy and the stability of targets tracking was improved.The experimental results showed that the method could achieve accurate detection and stable tracking in realtime.

关 键 词：目标探测 跟踪方法 帧间差分 水平集 直方图匹配 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]