Pansharpening by interspectral similarity and edge information using improved deep residual network

Peng Yu Chen, Shen-Chuan Tai

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Remote sensing image pansharpening involves the fusing of multispectral (MS) images with a panchromatic (PAN) image to produce an image with high-spatial as well as high-spectral resolution. We propose an improved pansharpening algorithm based on deep learning. A four-layer residual network is used as a reconstructed model to enable the accurate estimation of high-frequency details. We consider two priors to take advantage of MS information. The first prior indicates interspectral similarity, wherein the relationship between high- and low-resolution PAN images is used in the estimation of high-resolution MS images. The second prior provides the location of edges and textures according to the gradient of the PAN image. Consistency in the spectral characteristic is used as the basis in creating a pretrained model with the aim of accelerating convergence. Multiple evaluation metrics were applied to simulated and real images in order to compare the efficacy of the proposed method with that of state-of-the-art image fusion methods.

Original languageEnglish
Article number033013
JournalJournal of Electronic Imaging
Volume27
Issue number3
DOIs
Publication statusPublished - 2018 May 1

Fingerprint

Image fusion
Spectral resolution
Optical resolving power
Image resolution
Remote sensing
Textures
high resolution
image resolution
spectral resolution
learning
remote sensing
textures
fusion
gradients
Deep learning
evaluation

All Science Journal Classification (ASJC) codes

  • Atomic and Molecular Physics, and Optics
  • Computer Science Applications
  • Electrical and Electronic Engineering

Cite this

@article{e609bc073ef1465a95b4704e6904f39d,
title = "Pansharpening by interspectral similarity and edge information using improved deep residual network",
abstract = "Remote sensing image pansharpening involves the fusing of multispectral (MS) images with a panchromatic (PAN) image to produce an image with high-spatial as well as high-spectral resolution. We propose an improved pansharpening algorithm based on deep learning. A four-layer residual network is used as a reconstructed model to enable the accurate estimation of high-frequency details. We consider two priors to take advantage of MS information. The first prior indicates interspectral similarity, wherein the relationship between high- and low-resolution PAN images is used in the estimation of high-resolution MS images. The second prior provides the location of edges and textures according to the gradient of the PAN image. Consistency in the spectral characteristic is used as the basis in creating a pretrained model with the aim of accelerating convergence. Multiple evaluation metrics were applied to simulated and real images in order to compare the efficacy of the proposed method with that of state-of-the-art image fusion methods.",
author = "Chen, {Peng Yu} and Shen-Chuan Tai",
year = "2018",
month = "5",
day = "1",
doi = "10.1117/1.JEI.27.3.033013",
language = "English",
volume = "27",
journal = "Journal of Electronic Imaging",
issn = "1017-9909",
publisher = "SPIE",
number = "3",

}

Pansharpening by interspectral similarity and edge information using improved deep residual network. / Chen, Peng Yu; Tai, Shen-Chuan.

In: Journal of Electronic Imaging, Vol. 27, No. 3, 033013, 01.05.2018.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Pansharpening by interspectral similarity and edge information using improved deep residual network

AU - Chen, Peng Yu

AU - Tai, Shen-Chuan

PY - 2018/5/1

Y1 - 2018/5/1

N2 - Remote sensing image pansharpening involves the fusing of multispectral (MS) images with a panchromatic (PAN) image to produce an image with high-spatial as well as high-spectral resolution. We propose an improved pansharpening algorithm based on deep learning. A four-layer residual network is used as a reconstructed model to enable the accurate estimation of high-frequency details. We consider two priors to take advantage of MS information. The first prior indicates interspectral similarity, wherein the relationship between high- and low-resolution PAN images is used in the estimation of high-resolution MS images. The second prior provides the location of edges and textures according to the gradient of the PAN image. Consistency in the spectral characteristic is used as the basis in creating a pretrained model with the aim of accelerating convergence. Multiple evaluation metrics were applied to simulated and real images in order to compare the efficacy of the proposed method with that of state-of-the-art image fusion methods.

AB - Remote sensing image pansharpening involves the fusing of multispectral (MS) images with a panchromatic (PAN) image to produce an image with high-spatial as well as high-spectral resolution. We propose an improved pansharpening algorithm based on deep learning. A four-layer residual network is used as a reconstructed model to enable the accurate estimation of high-frequency details. We consider two priors to take advantage of MS information. The first prior indicates interspectral similarity, wherein the relationship between high- and low-resolution PAN images is used in the estimation of high-resolution MS images. The second prior provides the location of edges and textures according to the gradient of the PAN image. Consistency in the spectral characteristic is used as the basis in creating a pretrained model with the aim of accelerating convergence. Multiple evaluation metrics were applied to simulated and real images in order to compare the efficacy of the proposed method with that of state-of-the-art image fusion methods.

UR - http://www.scopus.com/inward/record.url?scp=85047720984&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85047720984&partnerID=8YFLogxK

U2 - 10.1117/1.JEI.27.3.033013

DO - 10.1117/1.JEI.27.3.033013

M3 - Article

AN - SCOPUS:85047720984

VL - 27

JO - Journal of Electronic Imaging

JF - Journal of Electronic Imaging

SN - 1017-9909

IS - 3

M1 - 033013

ER -