Remote sensing image scene classification method based on space and multi-channel fusion self-attention network

Abstract

The invention relates to the technical field of video image processing, and provides a remote sensing image scene classification method based on a space and multi-channel fusion attention mechanism, and the method comprises the steps: a, extracting the residual feature information of a remote sensing image through employing a ResNet network; b, performing feature mapping on the foreground content and the background content by using a spatial self-attention network to obtain spatial mapping features; b, performing multi-channel and multi-scale feature mapping on the residual feature information by using a multi-channel fusion self-attention network to obtain multi-channel fusion mapping features; c, synthesizing the extracted space mapping features and the multi-channel fusion mapping features; and d, classifying the synthesized mapping features by using a width classifier to obtain a classification result. The remote sensing image scene classification method based on space and multi-channel fusion provided by the embodiment of the invention can effectively improve the average precision of a remote sensing scene classification data set and reduce the calculation cost.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a scene classification method for remote sensing images. Background technique [0002] The purpose of remote sensing image scene classification is to assign a specific semantic category to remote sensing images. Remote sensing image scene classification technology has attracted much attention because of its application potential in urban monitoring, environmental detection, geographical structure analysis and other fields. The network framework for these tasks generally consists of two basic networks, namely feature mapping network and category classification network. [0003] In recent years, due to the powerful feature learning ability of deep convolutional neural networks (CNN), the accuracy of remote sensing image scene classification tasks has been significantly improved. The existing methods commonly used in remote sensing image scene classification tasks can be ...

AI Technical Summary

Problems solved by technology

First, to address the misidentification problem caused by incomplete feature representations, a new algorithm is provided, which enhances the relationship between feature maps across scales through a multi-channel fused attention module.

Images