Remote sensing image semantic segmentation method and device and storage medium

Abstract

The invention provides a remote sensing image semantic segmentation method, a remote sensing image semantic segmentation device and a storage medium, which are used for solving the technical problem of low accuracy of remote sensing image semantic segmentation. The device provided by the invention comprises a preprocessing module and a coding and decoding convolutional neural network. The input of the coding and decoding convolutional neural network model not only comprises the existing data channel, but also comprises a plurality of normalized indexes, so that the segmentation effect on a specific category is enhanced. In the coding path and the decoding path of the model, a plurality of groups of mixed extraction modules comprising a multi-scale convolution module are adopted to obtain feature information of different receptive fields, so that the segmentation effect on difficult points such as the same object and different spectrums and the same spectrum and foreign matters is enhanced. In the hybrid extraction module, an attention mechanism between channels is also adopted to enhance feature fusion between feature channels, and the feature extraction capability is further enhanced. The accuracy of semantic segmentation of the remote sensing image is improved, and the method can be applied to the field of semantic segmentation of various remote sensing images.

Description

technical field [0001] The present disclosure relates to the technical field of remote sensing image processing, in particular to a remote sensing image semantic segmentation method, device and storage medium. Background technique [0002] Remote sensing images refer to photos that record the size of electromagnetic waves of various ground features, mainly divided into aerial photos and satellite photos. The most widely used remote sensing images currently come from multi-band remote sensing, that is, the continuous electromagnetic spectrum with a wide range of ground radiation is divided into several narrower bands, and the same target is obtained at the same time by means of multi-band photography or multi-channel scanning. Remote sensing technology for spectral information in different bands. [0003] At present, the commonly used hyperspectral remote sensing image data can be divided into 4-band, 6-band, 7-band and other types according to the different imaging equipmen...

AI Technical Summary

Problems solved by technology

[0014] In view of this, the present disclosure provides a remote sensing image semantic segmentation method, device and storage medium for solving the technical problem of low accuracy of remote sensing image semantic segmentation

Images