摘要
模板匹配技术是运动目标跟踪中常用的方法。在基于差值测度的模板匹配算法中,满足匹配准则的最佳匹配点与真实位置偏差较大,这在运动目标跟踪中,由于偏差的不断累加会造成跟踪不稳定,甚至目标的丢失。针对这一问题,提出了结合目标模板内像素灰度值的空间分布情况,不断地对模板进行修正的算法,使得跟踪的实时性和稳定性得以提高。仿真实验结果证明了该算法的有效性。
Template matching is a method widely used in the area of target detection and tracking.The optimum matching location is different from the real location,which is computed in the traditional template matching algorithm.The difference is accumulated,which will lead to instability in moving target tracking,even losing the target being tracked.Therefore,an improved algorithm using spatial information of pixels in the target template is proposed.The algorithm can modify the template in real-time and the stability of moving target tracking is improved.Experimental result indicates the effectiveness of the method.
出处
《红外与激光工程》
EI
CSCD
北大核心
2007年第z2期197-200,共4页
Infrared and Laser Engineering
关键词
模板匹配
图像序列
搜索区域
Image sequence
Template matching
Search area
