Automatic Scene Extraction and Classification Method of Remote Sensing Image Based on Convolutional Neural Network

Abstract

The present invention claims a method for automatically extracting and classifying remote sensing image scenes based on a convolutional neural network. The main innovation is to modify the input feature map of the deep neural network in general and increase the data dimension. It mainly solves the problem that the resolution of remote sensing images is gradually increasing. For the classification of some confusing scenes such as residential areas, parks, roads, etc., it may appear that due to the limitation of the number of training samples, it is impossible to effectively extract features in further subdivision. For example, subdivide sparse residential areas and parks, dense residential areas and dense commercial areas in residential areas. The present invention uses deep learning or other methods to classify the remote sensing images to be classified pixel by pixel, and uses the classified result thematic map as a part of the feature map and the original scene map to be spliced ​​into the neural network for training and classification, so that the neural network It can fully learn the scene features and improve the accuracy of classification and recognition.

Description

technical field [0001] The invention belongs to the field of remote sensing image scene classification. On the basis of the traditional automatic classification of remote sensing image scenes, additional information is added and the dimension of information is expanded so that the neural network can more fully learn the differences in the characteristics of the scenes to be classified, so as to improve the accuracy of classification. Background technique [0002] (1) Convolutional neural network [0003] Convolutional Neural Network (CNN) is a feedforward neural network. Its artificial neurons can respond to surrounding units within a part of the coverage area, and it has excellent performance for large-scale image processing. It includes convolutional layers and pooling layers. [0004] Convolutional neural network is an efficient recognition method that has been developed in recent years and has attracted widespread attention. In the 1960s, Hubel and Wiesel found that i...

AI Technical Summary

Problems solved by technology

In general, the number of commonly used data sets or self-made data sets is not enough to allow the neural network to fully learn and avoid over-fitting, resulting in an accuracy that cannot meet expectations.

Images