基于硬件加速的空间碎片监视处理方法  

A Real Time Preprocessing Method for Hybrid Heterogeneous In-Orbit Meteorological Satellites

作　　者：蒋孝勇 曲耀斌[1] 周军[1] 刘波[1] 汪少林[1] 莫凡[1] 陈议[1] 桑峰[1] JIANG Xiaoyong;QU Yaobin;ZHOU Jun;LIU Bo;WANG Shaolin;MO Fan;CHEN Yi;SANG Feng(Shanghai Institute of Satellite Engineering,Shanghai 201109,China)

机构地区：[1]上海卫星工程研究所,上海201109

出　　处：《上海航天(中英文)》2024年第S2期175-182,共8页Aerospace Shanghai(Chinese&English)

摘　　要：为解决空间碎片识别难,空间碎片轨迹维护难问题,本文提出一种混合异构在轨目标实时处理平台技术。混合异构在轨实时处理方法可将检测结果直接下传地面用户,提高了空间碎片检测时效性。检测算法采用图像增强滤波手段,在轨实时滑窗滤波处理器为FPGA,滤波图像增强后的图像数据经过CPU过门限后得到空间碎片目标,空间碎片轨迹库由星载CPU管理维护。混合异构星载计算平台通过并行计算实现在轨实时处理。混合异构硬件加速平台可将传统单一CPU处理耗时9989.69 ms,提升到118 ms,达到图像帧250 ms实时处理能力,并实现目标轨迹稳定跟踪。To solve the problems of difficulty in identifying space debris and maintaining space debris trajectories,this paper proposes a hybrid heterogeneous real-time processing platform technology for in orbit targets.The hybrid heterogeneous in orbit real-time processing method can directly transmit the detection results to ground users,improving the timeliness of space debris detection.The detection algorithm adopts image enhancement filtering method,and the real-time sliding window filtering processor in orbit is FPGA.The image data after filtering and image enhancement passes through the CPU threshold to obtain the space debris target.The space debris trajectory library is managed and maintained by the onboard CPU.The hybrid heterogeneous onboard computing platform achieves realtime processing in orbit through image filtering and parallel computing.The processing time of a traditional single CPU is increased from 9989.69 ms to 118 ms.

关 键 词：空间碎片 混合异构 目标检测 硬件加速 在轨编程 

分 类 号：U666.1[交通运输工程—船舶及航道工程]