遥感星群面向密集目标的聚类规划方法  

Cluster planning method for dense targets by remote sensing satellite swarm

作　　者：杨玉恒 李小鹏 蔡恒明 李嘉敏 吴限德[1] YANG YuHeng;LI XiaoPeng;CAI HengMing;LI JiaMin;WU XianDe(College of Aerospace and Civil Engineering,Harbin Engineering University,Harbin 150001,China;Aviation University of Air Force,Harbin 150046,China)

机构地区：[1]哈尔滨工程大学航天与建筑工程学院,哈尔滨150001 [2]中国人民解放军空军航空大学,哈尔滨150046

出　　处：《中国科学:物理学、力学、天文学》2025年第2期70-82,共13页Scientia Sinica Physica,Mechanica & Astronomica

基　　金：国家自然科学基金(编号: 11772185)资助。

摘　　要：为了解决遥感星群对密集目标观测的效率低下、任务冲突、资源浪费等问题,首先提出了一种区域聚类方法,将互相接近的点目标整合为区域;然后按照卫星轨迹分割为条带,使得卫星单次成像可覆盖多个目标;随后对剩余的目标采用一种单景聚类方法,通过修改载荷指向位置使得卫星瞬时视场能够包含更多目标.使用聚类后的条带和指向点替换原始目标集合进行规划,对指向点采用凝视成像,对条带采用推扫成像.多组仿真算例表明,聚类规划能够有效降低任务数量,缓解目标观测冲突,减少计算耗时,且目标密集程度越高,上述效果愈加明显.In order to solve the problems of low efficiency, task conflicts, and resource waste in observing dense targets by remotesensing satellite swarm, this paper first proposes a regional clustering method, which integrates close point targets intoregions and then divides them into strips according to satellite trajectories, so that the satellite can cover multiple targetsin a single imaging. Subsequently, a mono scene clustering method is adopted for the remaining targets, which modifiedthe payload pointing position to include more targets in the instantaneous field of view. The original target set is replacedby the cluster strips and points, and the points are mapped by gaze imaging and the strips are mapped by push scanimaging. Simulation examples show that clustering planning method can effectively reduce the number of missions,alleviate target observation conflicts, and reduce computational time. Moreover, the higher the target density, the moreobvious the above effects become.

关 键 词：遥感星群 对地观测 任务规划 需求筹划 目标聚类 推扫成像 

分 类 号：P237[天文地球—摄影测量与遥感] TP751[天文地球—测绘科学与技术] TP311.13[自动化与计算机技术—检测技术与自动化装置]