Abstract: In order to explore dimensionality reduction and classification in hyperspectral remote sensing image, an algo- rithm based on marginal Fisher analysis(MFA) and k-nearest neighbor simplex(kNNS) was proposed in this paper. First,the data were projected from a high-dimensional space onto low-dimensional space by MF八combined with the in- formation of different classes. I}hen, classification was performed under the kNNS classifier by using a few neighbors from each class. The experimental results on the Urban data set, Washington DC Mall data set and Indian Pine hyper- spectral data set show the effectiveness of the proposed algorithm. When i(i=4,6,8) samples of each class arc randomly selected for training and 100 samples of each class for testing, the overall accuracy of our proposed algorithm is im- proved by 3. 7%一8. 5 0 o compared with other methods.

Key words: Hypcrspcctral images, Land cover classification, Graph embedding framework, Ncarcst neighbor