高光谱遥感影像混合像元分解定位联合模型研究

高光谱遥感影像亚像元定位技术能够确定混合像元中不同地物的空间分布。然而，传统亚像元定位方法大多基于丰度影像，其精度受限于混合像元分解结果的精度，从而限制了亚像元定位的适用性。本项目拟集成混合像元分解与亚像元定位技术，以“分解定位联合模型构建-空间特征自适应建模-联合模型快速求解”为研究主线，发展统一、稳健、快速的高光谱影像混合像元分解定位联合模型，揭示地物在亚像元中的空间分布信息，其最终成果可以提升高光谱遥感影像的亚像元信息处理能力与应用潜力，具有重要的理论与应用意义。

Sub-pixel mapping techniques for hyperspectral remotely sensed imagery can determine the spatial distribution of different land covers in mixed pixels. Most traditional sub-pixel mapping methods utilize the abundance map and the accuracies are limited by spectral unmixing techniques. This proposal develops a unified sub-pixel mapping model for hyperspectral imagery incorporating spectral unmixing techniques, which contains the following contents: construction of a unified model of spectral unmixing and sub-pixel mapping, adaptive modeling of the spatial features and fast solution of the unified model. The study can reveal the sub-pixel distribution of different land covers, and the sub-pixel information of hyperspectral remotely sensed imagery can be further utilized for application.