遥感影像稀疏表征与融合处理综合集成演示验证

Based on the verification platform of deployed integrated projects and theoretical and technical achievements on the key scientific problem of sparse representation and fusion processing in spatial information net, this project focuses on the comprehensive verification of sparse representation and fusion processing of remote sensing images in the space information network environment. According to the idea of "satellite-to-ground collaborative verification" at different stages and levels, the following contents are studied concretely: (1) verification of ultra-high precision attitude determination for small and light small star camera, (2) verification of intelligent target detection based on deep sparse representation, (3) verification of multi-temporal image change information extraction based on sparse computing, (4) verification of high-ratio intelligent compression based on image sparse representation, (5) integrated demonstration based on typical application scenarios. The aim of the project is to solve the bottleneck problems of high accuracy real-time processing on-board, intelligent extraction of target and change information, and task oriented high-ratio intelligent compression, so that it can provide support for the high precision intelligent real-time terminal application service of "fast", "precise " and " smart " for the massive remote sensing images under the limited bandwidth of space information network.

针对空间信息网络中空间信息稀疏表征与融合处理关键科学问题取得的理论方法成果，基于已部署集成项目构建的验证平台，本项目重点开展空间信息网络环境下的遥感影像稀疏表征与融合处理综合验证研究。本项目采用分阶段、分层次的星地协同验证思路，具体研究如下内容：（1）轻小型星相机超高精度姿态确定验证；（2）基于深度稀疏表征的智能目标检测验证；（3）基于稀疏计算的多时相影像变化信息提取验证；（4）基于稀疏表征的影像高倍智能压缩验证；（5）基于典型的应用场景的综合集成试验验证。基于上述研究，解决高精度在轨实时处理、目标和变化信息的智能提取以及面向任务的高倍率智能压缩等瓶颈难题，为空间信息网络有限带宽资源条件下，海量遥感影像“快”、“准”、“灵”的高精度智能实时终端应用服务提供支撑。

高分辨率卫星遥感影像具有数据海量、类型多样、信息丰富的特点，是空间信息网络的核心信息来源。通过稀疏表征与融合处理，一方面可提取有效空间信息，另一方面可降低传输带宽的压力，是实现遥感卫星快速、准确、灵活信息服务的关键。本项目重点围绕天基系统的静态和动态影像实时获取、在轨稀疏表征与智能处理及压缩传输取得的理论成果和突破的关键技术开展综合集成演示验证。主要针对海量遥感信息获取和应用终端遥感信息实时精准智能服务面临的高精度在轨实时处理、目标和变化信息的智能提取以及面向任务的高倍率智能压缩等瓶颈难题，以“珞珈三号01星”智能遥感卫星为主要研究平台，开展了轻小型星载相机超高精度姿态确定、动态目标实时智能检测与变化信息提取、基于影像稀疏表征面向任务的高倍影像智能压缩关键技术研究，以及基于智能遥感试验卫星的遥感影像稀疏表征与融合处理综合集成演示验证；实现了全球范围的信息获取、智能处理、稀疏表征与压缩传输和移动终端实时分发的全过程，为有限带宽条件下实现海量遥感影像“快”、“准”、“灵”的智能终端实时服务提供理论支撑和演示验证平台。在本项目支持下发表相关学术论文11篇，其中SCI核心期刊论文9篇；EI论文1篇；申请发明专利7项。本项目的研究成果、人才培养为我国空间信息网络的构建，提供了理论和技术支撑以及人才基础。