深空探测作为一个国家综合国力和科学技术发展水平的重要特征与标志，已引起世界各国的极大关注，拥有深空探测能力具有难以估量的潜在社会价值和军事意义。为保证我国深空探测计划的顺利推进，迫切需要开展深空目标形貌表征与提取识别的理论和技术的研究，为我国未来深空态势感知任务奠定技术基础，进一步加强我国进入、控制和利用深空的能力。本项目对深空目标表面形貌精细重建理论和技术开展研究，重点对光学传感器在线标定、多传感器数据联合区域网平差、多视影像密集匹配以及基于嵌入式系统的硬件加速等核心技术进行研究，研制深空目标表面形貌重建试验系统，为我国未来开展深空目标观测、抵近与着陆探测提供技术支撑，为精确、密集、快速地获取深空目标目标形貌提供技术储备。

Deep space exploration is important to one country, which reflects comprehensive national strength. Considering the current techniques of China's deep space exploration and foreign deep space exploration plan, it is so necessary to research topographic and morphological expression and extraction techniques of deep space objects surface, which is the base of future deep space situation exploration mission of China. In this project, theory and techniques of fine reconstruction of deep space object surface will be researched, which includes some key techniques such as calibration of optical camera in orbit, combined block adjustment with multi-sensor data, multi-view dense image matching and hardware acceleration based on embedded system. Finally an experimental system of reconstruction of deep space object surface will be developed based on these techniques, which is helpful to future deep space objects observation, arrival and landing missions. The aim of our research is producing digital surface models (DSM) of deep space objects precisely, densely and quickly.