A Remote Sensing Image Classification Method Using Color and Texture Feature

Abstract

In this paper, we present an effective remote sensing image classification method using color and texture feature based on D-S evidence theory and neural networks. In our method, the multiresolution Gabor filtering and the color components in PCA color space are applied. Firstly, PCA techniques are applied to RGB values of the original image. We apply components of the image besides the first principal component to train and classify the image using B-P neural network, then, we obtain a classification result. secondly, the texture images can be classified in multiple scales and orientations using the Gabor filtering, then, we obtain the second classification result. Finally, the two classification results of the B-P neural network are fused with evidence theory. The fused result is regarded as the final classification result of the original image. The experimental results show that the new method is efficient and improves the classification accuracy largely.