A Spatiotemporal Fusion Method Based on Multiscale Feature Extraction and Spatial Channel Attention Mechanism

Dajiang Lei,Liping Zhang,Weisheng Li,Gangsheng Ran

doi:10.3390/rs14030461

Dajiang Lei, Liping Zhang + Show 2 more

Open Access

https://doi.org/10.3390/rs14030461

Copy DOI

Abstract

Remote sensing satellite images with a high spatial and temporal resolution play a crucial role in Earth science applications. However, due to technology and cost constraints, it is difficult for a single satellite to achieve both a high spatial resolution and high temporal resolution. The spatiotemporal fusion method is a cost-effective solution for generating a dense temporal data resolution with a high spatial resolution. In recent years, spatiotemporal image fusion based on deep learning has received wide attention. In this article, a spatiotemporal fusion method based on multiscale feature extraction and a spatial channel attention mechanism is proposed. Firstly, the method uses a multiscale mechanism to fully utilize the structural features in the images. Then a novel attention mechanism is used to capture both spatial and channel information; finally, the rich features and spatial and channel information are used to fuse the images. Experimental results obtained from two datasets show that the proposed method outperforms existing fusion methods in both subjective and objective evaluations.

Highlights

With the development and progress of sensor technology, applications of remote sensing (RS) images in scientific research and human activities have become increasingly extensive [1,2]
The acquisition time of the reference image is denoted as T0, while the acquisition time of the predicted image is denoted as T1
(Moderate Resolution Imaging Spectroradiometer) image is denoted as M0, the Landsat image at T1 is denoted as L1, and the corresponding MODIS image is denoted as M1

Summary

Introduction

With the development and progress of sensor technology, applications of remote sensing (RS) images in scientific research and human activities have become increasingly extensive [1,2]. Some research areas and applications require RS images with a high temporal and spatial resolution. The term ’spatiotemporal fusion algorithm’ refers to the algorithmic fusion of at least two data sources with similar spectral ranges to generate data with more information than the original data sources [10,11]. These spatiotemporal fusion algorithms have been proven to be cost effective and useful [7]

Methods

Discussion

Conclusion