Dimensionality Reduction of Hyperspectral Image with Graph-Based Discriminant Analysis Considering Spectral Similarity

Remote Sensing

Fubiao Feng¹  Wei Li¹  Qian Du²  Bing Zhang³ 

[1] College of Information Science and Technology, Beijing University of Chemical Technology, Beijing 100029, China;Department of Electrical and Computer Engineering, Mississippi State University, Starkville, MS 39762, USA;Institute of Remote Sensing and Digital Earth, Chinese Academy of Science, Beijing 100094, China;

关键词: hyperspectral data;    dimensionality reduction;    graph embedding;    spectral similarity;   

DOI  :  10.3390/rs9040323

来源: DOAJ

【 摘 要 】

Recently, graph embedding has drawn great attention for dimensionality reduction in hyperspectral imagery. For example, locality preserving projection (LPP) utilizes typical Euclidean distance in a heat kernel to create an affinity matrix and projects the high-dimensional data into a lower-dimensional space. However, the Euclidean distance is not sufficiently correlated with intrinsic spectral variation of a material, which may result in inappropriate graph representation. In this work, a graph-based discriminant analysis with spectral similarity (denoted as GDA-SS) measurement is proposed, which fully considers curves changing description among spectral bands. Experimental results based on real hyperspectral images demonstrate that the proposed method is superior to traditional methods, such as supervised LPP, and the state-of-the-art sparse graph-based discriminant analysis (SGDA).

【 授权许可】

Unknown