期刊名称:International Journal on Smart Sensing and Intelligent Systems
印刷版ISSN:1178-5608
出版年度:2016
卷号:9
期号:4
页码:1991-2009
出版社:Massey University
摘要:Hyperspectral data has rich spectrum information, strong correlation between bands andhigh data redundancy. Feature band extraction of hyperspectral data is a prerequisite and an importantbasis for the subsequent study of classification and target recognition. Deep belief network is a kind ofdeep learning model, the paper proposed a deep belief network to realize the characteristics bandextraction of hyperspectral data, and use the advantages of unsupervised and supervised learning ofdeep belief network, and to extract feature bands of spectral data from low level to high-level gradually.The extracted feature band has a stronger discriminant performance, so that it can better to classifyhyperspectral data. Finally, the AVIRIS data is used to extract the feature band, and the SVM classifieris used to classify the data, which verifies the effectiveness of the method.
关键词:Hyperspectral; Feature extraction; Deep learning; Deep belief network; Restricted boltzman;machines