深空可见光图像中弱小运动目标实时检测  

Real Time Detection for Dim Moving Target in Visible Optical Image of Deep Outer Space

作　　者：柳庆武[1,2] 胡晓惠[1] 袁麟[1] 

机构地区：[1]中国科学院软件研究所综合信息系统技术国家级重点实验室,北京100190 [2]中国科学院研究生院,北京100190

出　　处：《电子学报》2009年第7期1614-1617,共4页Acta Electronica Sinica

基　　金：中科院知识创新工程(No.KGCX3-SYW-407-03)

摘　　要：针对深空可见光图像背景特征与目标特性,提出了迭代质心的方法自适应地搜索恒星灰度质心作为特征点并构造基于恒星空间分布的特征模型,实现亚像素级精度的图像序列配准,通过8-邻域联通聚类分析的路径判别法,解决了弱小运动目标实时检测.Through the analysis of characteristics about background and target in visible optical image of deep outer space,the matching of image sequence at sub-pixel is realized by the method of iterative computing which is proposed to adaptive searching the pixel-mass center of fixed star and the mtxiel of characteristics which is constructed on spatial distribution. Based on image sequence matching and dividing,the arithmetic of real-time detection for dim moving target is designed by the clustering on 8-neighbour-pixel.

关 键 词：深空可见光图像 弱小运动目标 实时检测 迭代质心 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]