高光谱遥感图像目标探测的新理论新方法及关键技术研究

高光谱遥感图像目标探测的研究在军事与民用上都具有重要的理论与应用价值，是目前遥感图像处理领域中的研究热点。如何实现高精度的目标探测，如何克服光谱变异性问题，如何有效实现非线性的目标探测，则是当前研究中的难点问题。. 本项目针对高光谱目标探测的难点问题，借助于稀疏表示、低秩分解、多目标优化、非线性光谱混合模型、核方法和流形学习理论，深入开展光谱匹配和光谱异常目标探测两个方面的研究。本项目将研究：1)空谱联合的稀疏表达与多目标优化；2)结合空间平滑性的高光谱图像低秩分解与多目标优化；3)克服光谱变异性问题的高光谱目标探测；4)基于非线性模型的非线性目标探测；5)不基于非线性模型的非线性目标探测。所提出的理论与方法将在实际高光谱数据上进行大量的实验验证。. 本项目将突破当前的高光谱目标探测理论框架，为实现地面目标的高精度检测与识别，提供新理论、新方法以及解决关键问题和实现关键技术的具体方案。

The research of target detection in hyperspectral remote sensing images has important theoretical and practical value in both military and civil fields. It is a hot spot in the field of remote sensing image processing at present. How to achieve high-precision target detection, how to overcome the spectral variability problem, and how to effectively realize nonlinear target detection are the difficult problems in current research.. This project aims at the difficult problems of hyperspectral target detection and deeply carries out the research on the two aspects of spectral-matched and spectral-anomalous target detections by means of sparse representation, low-rank decomposition, multi-objective optimization, nonlinear spectral mixing models, kernel method and manifold learning theories. This project will study: (1) spatial-spectral information based sparse representation with multi-objective optimization; (2) spatial smoothness based low-rank decomposition for hyperspectral images with multi-objective optimization; (3) overcoming the problem of spectral variability for hyperspectral target detection; (4) nonlinear target detection based on nonlinear models; (5) nonlinear target detection without nonlinear models. The proposed theories and methods will be verified by a large number of experiments on the actual hyperspectral data.. This project will break through the current theoretical framework of hyperspectral target detection and provide novel theories, novel methods and concrete schemes to solve key problems and realize key technologies for achieving high-precision detection and identification of ground targets.