中文摘要：

在高光谱图像目标检测领域中,稀疏表示算法取得了较好的检测效果,但传统基于稀疏表示的目标检测算法稀疏向量的求解耗时长,检测时只利用高光谱图像的光谱信息,没有考虑空间信息,且其字典中所包含训练样本种类和数目较少,都对目标检测有一定影响。针对上述不足,通过对字典进行改进,添加空间信息,转变稀疏向量求解思路,提出基于稀疏表示的高光谱图像增殖快速目标检测算法。通过实验仿真证明,此算法在目标检测精度上有一定程度的提高,并且缩短了算法的运算时间。

英文摘要：

Sparse representation algorithm has been successfully applied in the field of hyperspectral imagery （HSI） target detection and achieved nice results. However, the traditional target detection algorithm takes a long time to solve the solution of sparse vector, which only uses the spectral information of the hyperspectral imagery instead of considering the spatial information, so there is lack of training samples in the traditional dictionary. They have some bad effects on the target detection. The proliferative fast algorithm based on sparse representation for hyperspectral imaging target detection （PFSR） has been proposed. The new algorithm improves the dictionary, adds the spatial information to the process of target detection and optimizs the ideas of solving sparse vector. The simulation results prove that the proliferative fast target detection algorithm does work and the new algorithm can not only improve the accuracy in target detection but also shorten the time on target detection.