摘要
针对在轨光学相机视轴无法通过卫星姿态进行预估,序列探测图像中恒星背景又无法忽略等空间运动点目标检测中的难点问题,提出了一种基于序列帧探测图像运动特性分析的点目标检测方法。该方法首先从恒星星图中检测出所有的目标点,接着选取亮星点进行恒星识别,最后对所有的目标点进行运动分析,检测出与恒星具有不同运动特性的星点作为空间点目标。仿真结果表明,该方法能够从运动的恒星背景中识别出空间运动点目标,对完善空间运动点目标检测方法有参考价值。
In order to solve the problems that the foresight of camera cannot be estimated by using attitude of satellite platform and star background cannot be ignored in sequence images,a spatial moving point target detection method was proposed based on analyzing motion features for on- orbit optical camera. This method firstly detects all the target points from star maps;secondly selects bright point targets to go through star recognition; and thirdly the movement of stars can be estimated from continuous image frames; finally spatial point targets can be detected by analyzing the differences of movement between point targets and stars. Through simulation experiments,the proposed method can identify spatial point target from moving star background in sequence images.
出处
《武汉理工大学学报(信息与管理工程版)》
CAS
2015年第6期667-670,共4页
Journal of Wuhan University of Technology:Information & Management Engineering
基金
国家自然科学基金(青年基金)资助项目(61503410)
关键词
在轨光学相机
空间点目标检测
星图识别
运动参数估计
on-orbit optical camera
spatial point target detection
star pattern recognition
motion parameter estimation