在轨实时空间目标检测算法研究  

Research on the On-orbit Real-time Space Target Detection Algorithm

作　　者：韩璐瑶 谭婵[2] 刘云猛[2] 宋锐[1] HAN Luyao;TAN Chan;LIU Yunmeng;SONG Rui(State Key Laboratory of Integrated Service Networks,Xidian University,Xi′an 710071,China;Shanghai Institute of Technical Physics of the Chinese Academy of Sciences,Shanghai 200083,China)

机构地区：[1]综合业务网理论及关键技术国家重点实验室,西安电子科技大学,西安710071 [2]中国科学院上海技术物理研究所,上海200083

出　　处：《航天返回与遥感》2021年第6期122-131,共10页Spacecraft Recovery & Remote Sensing

摘　　要：空间目标监视是保证空间活动正常运行的重要前提,利用光学设备进行光电探测是空间目标监视的重要手段之一。在天文光电深空图像序列中,恒星目标和空间目标成像特性相似,对目标检测产生干扰,因此实时稳健地进行空间目标检测有着很大的难度。文章对天文光电图像目标成像的规律进行了分析研究,提出一种基于帧间匹配的目标检测算法:首先通过形态学操作进行图像预处理,再通过帧间匹配提取特征,进而检测空间目标。FPGA与DSP硬件平台下的图像数据测试结果显示,该算法对不同工况均具有良好的检测概率和较低的虚警概率,图像处理速度可达5帧/s,具备有效性和实时性,适用于深空背景下天文光电探测系统对空间目标的实时检测需求。Space target surveillance is an important prerequisite for ensuring the normal operation of space activities. The use of optical equipment for photoelectric detection is one of the important means of space target surveillance. In the astronomical photoelectric deep-space image sequence, the imaging characteristics of stellar targets and space targets are similar, which will interfere with target detection. Real-time and robust target detection is also very difficult. The article analyzed and studied the astronomical photoelectric image target imaging law, and proposed a target detection algorithm based on frame matching. By performing image preprocessing through morphological operations, and extracting features through inter-frame matching, and then detect spatial objects were detected. This algorithm is suitable for the real-time detection requirements of the astronomical photoelectric detection system for space targets under the background of deep space. Under the experiment of FPGA and DSP hardware platform, the validity and real-time performance of the algorithm were verified.

关 键 词：目标检测 形态学 实时图像处理 可见光图像 空间遥感 

分 类 号：TN957.51[电子电信—信号与信息处理]