摘要
随着航天技术的发展,太空目标精细感知在动态监测、形变分析和轨道交会中具有重要作用.极端空间环境下的图像退化、高动态场景及多模态融合问题限制了感知精度和效率.本文聚焦极端空间环境目标精细感知的3个核心问题,包括低质图像增强、目标三维重建和目标姿态估计,分析其研究进展与挑战.针对低质图像增强,本文回顾了去噪、去模糊及超分辨率等领域经典方法,讨论了不同方法应对极端光照、各类噪声等干扰的解决思路和优缺点.针对目标三维重建,结合光学重建与点云配准重建技术,探讨动态场景下的优化策略.针对目标姿态估计,分析了高动态环境下的传感器可靠性及高精度推断方法.本文呈现了领域方法发展时间脉络图、方法分类图及数据库与评价指标表,为极端空间感知研究提供全面总结和参考,讨论了未来需要突破的瓶颈问题,助力极端空间环境精细感知新技术的发展.
With the advancement of space technology,precise perception of space objects plays a vital role in dynamic monitoring,deformation analysis,and orbital rendezvous.However,challenges such as image degradation in extreme space environments,highly dynamic scenes,and multimodal fusion issues limit the accuracy and efficiency of perception.This paper focuses on three core issues in the precise perception of fine-grained objects in extreme space environments:low-quality image enhancement,3D reconstruction of objects,and object pose estimation,analyzing their research progress and challenges.For low-quality image enhancement,we review classical methods in denoising,deblurring,and super-resolution,discussing the approaches and advantages and disadvantages of various methods in addressing extreme lighting conditions,various types of noise,and other interferences.For 3D object reconstruction,we explore optimization strategies in dynamic scenes by integrating optical reconstruction with point cloud registration technologies.For object pose estimation,we analyze sensor reliability and high-precision inference methods in highly dynamic environments.The paper presents a chronological timeline of method development,a classification diagram of methodologies,and tables summarizing databases and evaluation metrics,providing a comprehensive overview and reference for research on extreme space perception.Furthermore,it discusses critical bottlenecks that need to be addressed in the future to advance new technologies for precise perception in extreme space environments.
作者
张楚艾
杨佳琪
闫庆森
齐召帅
邹阳
张艳宁
Chu’ai ZHANG;Jiaqi YANG;Qingsen YAN;Zhaoshuai QI;Yang ZOU;Yanning ZHANG(School of Computer Science,Northwestern Polytechnical University,Xi’an 710129,China)
出处
《中国科学:信息科学》
北大核心
2025年第9期2193-2249,共57页
Scientia Sinica(Informationis)
基金
国家自然科学基金(批准号:62372377,62301432,62306240,62005217)
陕西省自然科学基础研究计划(批准号:2023-JC-QN-0685,QCYRCXM-2023-057)
中央高校基本科研业务费专项资金资助项目。
关键词
太空感知
低质图像增强
三维重建
姿态估计
极端环境
space perception
low-quality image enhancement
3D reconstruction
pose estimation
extreme environments
