空间非合作目标航天器太阳帆板识别方法研究

Research on solar panel recognition method of space non-cooperative target spacecraft

面向非合作航天器相对位姿测量需求,提出一种无需先验信息和人工干预的航天器太阳帆板检测算法。在分析太阳帆板视觉特征基础上,设计了三级分类器级联的目标检测算法。针对角度未知长矩形目标识别方法独立性和稳健性不强的问题,设计了旋转图片检测目标的方法;针对传统特征提取方法效率低的问题,提出了颜色分布积分图和基于积分图的特征提取方法。最后通过实验验证了在不同背景下太阳帆板检测任务中,提出算法的实用性和鲁棒性。

Aiming at the relative pose measurement requirements of non-cooperative satellite targets,a spacecraft solar panel detection algorithm without prior information and manual intervention is proposed.Based on the analysis of the visual characteristics of solar windsurfing,a three-level classifier cascaded target detection algorithm is designed.Aiming at the problem of the lack of independence and robustness of the long rectangular target recognition method with unknown angle,a method for detecting targets by rotating pictures is designed.Aiming at the problem of low efficiency of traditional feature extraction methods,color distribution integral maps and feature extraction methods based on integral maps are proposed.Finally,it is verified through experiments that the proposed algorithm has good practicability and robustness in the solar sail detection task under different backgrounds.