Fast processing method of high resolution remote sensing image based on decision tree classification

Abstract

In order to accurately distinguish the information of the nodes to be identified in remote sensing images and generate a global image processing strategy, a fast processing method for high-resolution remote sensing images based on decision tree classification is proposed. According to the classification principle of decision tree organization, the image data is preprocessed to complete the UAV remote sensing image coordination and registration, and the remote sensing image edge detection based on decision tree classification method is realized. On this basis, the convolution neural network is established, and the regularized constraint processing results are obtained by predicting the image scale. The fast processing method of high-resolution remote sensing image based on decision tree classification is successfully applied. The experimental results show that, compared with the traditional feature point matching principle, the average detection accuracy of the fast processing method is higher, and the node parameter matching accuracy is higher, which meets the practical application requirements of accurate resolution of the information to be identified in remote sensing images.