摘要
提出了一种用于高速目标跟踪的快速识别算法.根据图像目标与背景的特点,摒弃传统的逐点扫描方法,设计了一种“十字”匹配模板,根据待识别区域与“十字”模板的灰度差进行目标识别;为了提高识别速度,对目标图像进行二值化处理,采用逻辑运算获得模板与目标的相似性测度,通过相似性测度与阈值的比较,进行遍历点的跳跃.仿真结果表明,该算法大大提高了识别速度,实现了对高速摄像机拍摄角度的控制,进而对目标进行实时跟踪.
A fast identification algorithm for high speed target is presented.According to the characteristics of image target and background,a cross matching template is designed which distinguishes the target by the difference of grayness between the area to be identified and the cross matching template.Thus the traditional point by point scanning method is abandoned.In order to improve the speed of identification,target images are binary processed.Logic calculation is adopted to obtain the comparability estimate between the matching template and target.By comparing the comparability estimate and its threshold,the scanning points skip.The simulation results showed that this algorithm improves the identification speed greatly and realizes the control of shooting angle for the high speed vidicon,so as to track the target real time.
出处
《北京理工大学学报》
EI
CAS
CSCD
北大核心
2006年第11期1026-1029,共4页
Transactions of Beijing Institute of Technology
基金
国家部委预研项目(4040508010501)
关键词
模板匹配
最大相似性
目标识别
目标跟踪
template matching
maximum comparability
target identification
target tracking
