一种恒星背景的空间目标检测方法  

A Method for the Detection of Artificial Target in Star Background

作　　者：庞嘉鸿 黄鹂[2] 明德烈[1] PANG Jiahong;HUANG Li;MING Delie(School of Artificial Intelligence and Automation,Huazhong University of Science and Technology,Wuhan 430074;Beijing Aerospace Automatic Control Institute,Beijing 100854)

机构地区：[1]华中科技大学人工智能与自动化学院,武汉430074 [2]北京航天自动控制研究所,北京100854

出　　处：《计算机与数字工程》2024年第10期3100-3106,共7页Computer & Digital Engineering

摘　　要：为解决在远距离探测阶段空间目标与恒星的光谱特性和运动相似导致难以检测的问题,研究首先提出多源标准星型模式的星图识别算法,构建标准星型模式,合并多个观测星的识别结果;再提出多帧特征融合判决方法,应用星图仿真、均值漂移聚类以及置信度判决。实验结果表明,对比传统拓扑结构的星图识别算法,多源标准星型模式对星点位置噪声和星等噪声具有更强的抗干扰能力,识别率均高于95%且时间少于2 s,多帧特征融合判决方法准确识别与恒星运动相似的空间目标。该研究提出的方法可以检测恒星背景的空间目标。To address the difficulty of the artificial target detection owing to the similarities of the spectral characteristics and motion of artificial target and stars,this study firstly proposes a star identification algorithm with multi-source standard star pattern.The algorithm establishes a standard star pattern and fused multi-star identification results.A multi-frame feature fusion decision method applied with mean-shift clustering and confidence measure is secondly proposed.The results show that compared with the traditional topology star identification algorithm,the multi-source standard star pattern is of stronger anti-interference capacity to the star location noise and magnitude noise.The identification rate is higher than 95%and the time is less than 2 s.The multi-star fusion identification algorithm could accurately identify artificial target from the stars with similar motion.The methods proposed in this study realizea the artificial target detection in star background.

关 键 词：空间目标 星图识别 标准星型模式 融合判决 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]