Pan-sharpening via multi-scale and multiple deep neural networks

作者：

Highlights：

• Fully extracting the PAN image spatial structure features.

• Multi-scale and multiply DNNs needs less training time than one DNN.

• The fusion time does not increase as the number of MS bands increases.

• High-quality images are the basis of interpretation of remote sensing images in HCI.

摘要

•Fully extracting the PAN image spatial structure features.•Multi-scale and multiply DNNs needs less training time than one DNN.•The fusion time does not increase as the number of MS bands increases.•High-quality images are the basis of interpretation of remote sensing images in HCI.

论文关键词：Deep neural network (DNN),Residual compensation,Multispectral image,Pan-sharpening

论文评审过程：Received 11 October 2019, Revised 11 March 2020, Accepted 25 March 2020, Available online 11 April 2020, Version of Record 14 April 2020.

论文官网地址：https://doi.org/10.1016/j.image.2020.115850