A novel hyperspectral remote sensing imagery feature extraction algorithm called discriminative supervised neighborhood preserving embedding (DSNPE ) is proposed for supervised linear feature extraction.DSNPE can preserve the local manifold structure and the neighborhood structure.What’s more, for each data point,DSNPE aims at pulling the neighboring points with the same class label towards it as near as possible,while simultaneously pushing the neighboring points with different labels away from it as far as possible.Numerical experiments in three real hyperspectral- image datasets are reported to illustrate the out performance of DSNPE when compare DSNPE with a few computing methods,such as PCA,NWFE, LPP and NPE.%提出一种名为判别监督邻域保留嵌入的新型特征提取算法(discri minative supervised neigh-borhood preserving embedding,DSNPE)。在高维超光谱遥感图像特征提取过程中,DSNPE 不但能保留图像的局部流形结构和邻域信息,而且采用像素点由邻域同类像素点线性表示,将邻域中同类和非同类像素点分开处理,利用判别分式求解最优投影矩阵,使高维像素点投影到低维空间时,同类点离得尽可能近,非同类点离得尽可能远,有利于图像的分类。对3幅超光谱遥感图像的特征提取及分类的试验说明,与主成分分析(PCA)、非参数权重特征提取(NWFE)、局部保留投影(LPP)、邻域保留嵌入(NPE)等相比,具有一定的优越性和可判别性。
展开▼
