Abstract: The application of remote sensing image classification has important significance in the
research of remote sensing image. In order to improve the classification accuracy of hyperspectral
remote sensing images, this paper proposed the hyperspectral classification method based on multiple
features fusion method. The method normalized and fused spatial feature and spectral feature, and then
use the AdaBoost ensemble algorithm to classify the multiple features. First, the method uses principal
component analysis to reduce the dimensionality of hyperspectral data, and extracts the texture feature
and the histogram feature, then the three features will be normalized, finally uses the AdaBoost
ensemble algorithm classification method to classify the remote sensing data of hyperspectral remote
sensing. The result of experimentals shows that compared to the single feature classification, this
proposed method can achieve higher classification accuracy and better classification performance.

Key words: image classification, multiple features, AdaBoost, ensemble algorithm, classification accuracy