一种序列CCD星图中运动目标识别的稳健算法  

Robust Algorithm for Moving Target Recognition of Sequential CCD Star Images

作　　者：胡敏[1] 曾国强[2] 

机构地区：[1]装备指挥技术学院研究生管理大队,北京101416 [2]装备指挥技术学院航天装备系,北京101416

出　　处：《装备指挥技术学院学报》2007年第5期69-72,共4页Journal of the Academy of Equipment Command & Technology

基　　金：国家安全重大基础研究资助项目

摘　　要：运动星图背景下运动目标识别方法研究在天基光学空间目标探测方面具有重要意义。针对CCD（charge coupled device）星图的特点,采用阈值分割算法对原始星图进行预处理,利用基于连通域的算法完成对星点位置提取;提出了利用星点间角度信息确定目标位置的算法;对分辨率为768×576的序列CCD实拍星图的实验研究表明了该算法的有效性和稳定性。The research on methods of moving target recognition of moving star images is significant for space based optical target detection. Aimed at the characteristics of star images, the threshold is determined by use of an adaptive segmentation algorithm. Then a new effective algorithm for the location extractions of star points based on the connection of pixels is represented. Based on this, the paper puts forward a new algorithm for locating the moving targets in the real-time star images by using of the relative angle information of the star points. Results of the experiment on the real-time star images with resolution of 768 × 576 grabbed by charge coupled device （CCD） indicate that the algorithm is robust and effective.

关 键 词：目标识别 图像处理 特征提取 

分 类 号：V412[航空宇航科学与技术—航空宇航推进理论与工程]