天基平台小目标特征提取算法研究  

Discussion on Space-Based Small Target Features Extraction Algorithm

作　　者：任超伟[1,2] 刘恩海[1] 钟杰[1] 

机构地区：[1]中国科学院光电技术研究所,成都610209 [2]中国科学院研究生院,北京100039

出　　处：《计算机与数字工程》2009年第2期11-13,19,共4页Computer & Digital Engineering

摘　　要：针对天基平台凝视模式下小目标成像的特点,根据目标的检测、识别、跟踪的实时性要求,提出了一种以线段表为基础的小目标重心、面积、平均灰度、形状因子等特征的提取算法,并在这些参数的基础上提出了以孤星为基础的图像配准算法提取出目标的配准参数。在Matlab环境下,利用星模拟器输出图像对算法作了验证。In view of space- based imaging characters on staring mode, to satisfy real time needing performance on target detecting, recognizing and tracking, a line table based algorithm was presented for extracting small target features such as cg, area, average brightness, shape factor. Besides, on the basis of these features, an idea was given for image registration by finding the isolate star to extracting the registration parameter. In the end of this article, the author verified the algorithm using the star simulator output image in the Matlab environment.

关 键 词：小目标 线段表 目标特征 图像配准 孤星 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]