High Resolution Remote Sensing Image Classification Algorithm Based on Improved FCN

Abstract

Aiming at the problems of less bands of high resolution remote sensing image data and limited learning richness of model features, this paper proposes a high resolution remote sensing image classification algorithm based on improved full convolution neural network. Firstly, a standardization layer is added to batch process the image, and then a pooling index is added to the image to realize the up-sampling. Finally, the pooling index, the transposed convolution and the convolution eigenvalue are combined into a feature group to restore the class pixels of the image to a great extent. It can improve the prediction ability of the model. A simulation experiment is carried out to verify the effectiveness and feasibility of the proposed algorithm.