微型航天器多目标捕获算法研究  

Algorithmic Research on Multi-target Capture for Microspacecraft

作　　者：邵秀娟[1,2] 胡柄樑[1] 闫鹏[1,2] 

机构地区：[1]中国科学院西安光机所光谱室,西安710068 [2]中国科学院研究生院,北京100039

出　　处：《微计算机信息》2010年第16期201-203,共3页Control & Automation

摘　　要：为了实时性捕获混杂在天基背景中的运动目标,采用一种基于Fourier-Mellin图像配准的双差分算法。先采用Fouri-er-Mellin算法和相位相关算法求取图像的旋转角度和缩放及平移参数,得出配准点,对连续采集的三帧图像中的前两帧和后两帧分别进行配准,差分,二值化,然后将两幅二值化图像逻辑乘,识别出目标。仿真结果表明该算法虚警率低,适合微型航天器实时多目标捕获的要求。In order to capture the targets which were intermixed in the space background, this paper proposed an algorithm that com- bines image registration based on Fourier-Mellin algorithm with double difference method. Firstly, we use Fourier-Mellin algorithm and phase correlation algorithm to get the rotation, scaling and translation parameters. The algorithm separately use image registration, image difference to the former two images and the later two images which were collected three seriate images, then multiply the two difference images by pixel after binarizing to detect the target. The result of simulation indicates that the recognized result is good. So the double difference based on image registration is relatively fit microspacecraft target detector to identify the spatial target.

关 键 词：微型航天器 Fourier—Mellin算法 图像配准 双差分 

分 类 号：V423.6[航空宇航科学与技术—飞行器设计]