图像相关匹配算法研究与红外目标跟踪应用  

Image Correlation Matching Algorithms and Infrared Target Tracking

作　　者：宿燕鸣[1] 齐敏[1] 李大健[1] 严欣[1] 

机构地区：[1]西北工业大学,陕西西安710129

出　　处：《计算机仿真》2011年第9期46-49,共4页Computer Simulation

基　　金：三维医学模型检索系统(Z2010044)

摘　　要：研究红外目标跟踪问题,针对目标的准确定位,图像相关匹配技术是目标跟踪中最基本的方法。当前红外成像导引头要求实时跟踪,但是在目标跟踪的末端,匹配点漂移和滑动将直接影响目标跟踪的精度。为提高定位精度和实时性,提出了一种新的实时跟踪算法。算法从红外图像中获取特征点,以特征点为中心选取参考模板,利用边缘检测算法获得边缘点集,使用自适应阈值的SSDA算法进行边缘点集的匹配,实现实时目标跟踪仿真。实验结果表明,算法很好地解决了红外目标跟踪精度问题,并满足实时性和跟踪稳定性要求。The image correlation matching has become a basic method of tracking, and real-time target tracking is the the key technology of imaging infrared seeker. In the end of tracking, the drifting and slideing in match points can directly affect the tracking accuracy. In this paper, a new real-time tracking algorithm is proposed based on image correlation matching algorithms and SUSAN edge detection algorithm. The algorithm gets the feature points from infra- red images, and takes feature points as the center of the reference template, and then gets the edge points set by edge detection algorithm. Finally, it uses adaptive threshold SSDA algorithm to match edge points set. Experimental results show that the algorithm solves the problems well in target tracking such as point sliding and drifting, and it also meets real time requirements and has a certain degree of stability.

关 键 词：相关匹配 目标跟踪 边缘检测 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]