稀疏表达模型在高光谱遥感影像目标探测中的应用

为了更好地利用高光谱影像的空间和光谱信息,提出了一种基于稀疏表达模型的高光谱遥感影像目标探测方法.首先通过对影像训练样本进行训练提取过完备字典,利用稀疏表达模型对遥感影像稀疏表达既达到降维的目的,又可以表示出遥感影像的主要信息;然后利用传统的目标探测器结合目标已知光谱信息对高光谱遥感影像进行目标探测,即基于稀疏表达模型的高光谱遥感影像目标探测(SRM-TD).3种影像数据的实验结果表明:在确定的迭代次数下,通过设置稀疏度L可以得到最优的探测结果.提出的探测方法在参数设置、选择和运行结果上优于传统的高光谱遥感影像目标探测方法.

Application of Sparse Representation Model in Target Detection of Hyperspectral Remote Sensing Image

In order to make better use of spatial and spectral information in hyperspectral imaging, a sparse representation model was presented based on high spectral imaging target detection method in this paper. First, by extracting the over-complete dictionary through training samples, the sparse representation model of remote sensing image sparse expression was established for dimensionality reduction purposes. And the main information of remote sensing image was presented. Then, the hyperspectral remote sensing image target detection ( SRM-TD) based on the sparse expression model was used to detect the hyperspectral remote sensing image by using the traditional target detector combined with the target known spectral information. The experimental results of the three kinds of image data show that the optimal detection result can be obtained by setting the degree of sparse L under the number of iterations. The proposed detection method is superior to the traditional high spectral imaging target detection method in the parameter setting, selecting and operating results.