Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks

Visible (VIS) imagery is important for monitoring tropical cyclones (TCs) but is unavailable at night. This study presents a conditional generative adversarial networks model to generate nighttime VIS imagery with significantly enhanced accuracy and spatial resolution. Our method offers three key im...

Full description

Saved in:
Bibliographic Details
Main Authors: Jinghuai Yao, Puyuan Du, Yucheng Zhao, Yubo Wang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10988561/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850269620632551424
author Jinghuai Yao
Puyuan Du
Yucheng Zhao
Yubo Wang
author_facet Jinghuai Yao
Puyuan Du
Yucheng Zhao
Yubo Wang
author_sort Jinghuai Yao
collection DOAJ
description Visible (VIS) imagery is important for monitoring tropical cyclones (TCs) but is unavailable at night. This study presents a conditional generative adversarial networks model to generate nighttime VIS imagery with significantly enhanced accuracy and spatial resolution. Our method offers three key improvements compared to existing models. First, we replaced the L1 loss in the pix2pix framework with the structural similarity index measure (SSIM) loss, which significantly reduced image blurriness. Second, we selected multispectral infrared bands as input based on a thorough examination of their spectral properties, providing essential physical information for accurate simulation. Third, we incorporated the direction parameters of the sun and the satellite, which addressed the dependence of VIS images on sunlight directions and enabled a much larger training set from continuous daytime data. The model was trained and validated using data from the advanced Himawari imager in the daytime, achieving statistical results of SSIM = 0.923 and root mean square error = 0.0299, which significantly surpasses existing models. We also performed a cross-satellite nighttime model validation using the day/night band of the visible/infrared imager radiometer suite, which yields outstanding results compared to existing models. Our model is operationally applied to generate accurate VIS imagery with arbitrary virtual sunlight directions, significantly contributing to the nighttime monitoring of various meteorological phenomena.
format Article
id doaj-art-4c0f5a72adbf4dfd9a55dcacb8aa831d
institution OA Journals
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-4c0f5a72adbf4dfd9a55dcacb8aa831d2025-08-20T01:53:04ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118126161263310.1109/JSTARS.2025.356707410988561Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial NetworksJinghuai Yao0https://orcid.org/0009-0002-0525-8222Puyuan Du1https://orcid.org/0009-0003-9088-9586Yucheng Zhao2https://orcid.org/0009-0006-5259-0868Yubo Wang3https://orcid.org/0009-0004-7125-9762Department of Astronomy, University of Wisconsin–Madison, Madison, WI, USADepartment of Chemistry and Biochemistry, University of California, Los Angeles, Los Angeles, CA, USASchool of Health Sciences, Guangzhou Xinhua University, Guangzhou, ChinaDepartment of Atmospheric and Oceanic Sciences, University of Wisconsin–Madison, Madison, WI, USAVisible (VIS) imagery is important for monitoring tropical cyclones (TCs) but is unavailable at night. This study presents a conditional generative adversarial networks model to generate nighttime VIS imagery with significantly enhanced accuracy and spatial resolution. Our method offers three key improvements compared to existing models. First, we replaced the L1 loss in the pix2pix framework with the structural similarity index measure (SSIM) loss, which significantly reduced image blurriness. Second, we selected multispectral infrared bands as input based on a thorough examination of their spectral properties, providing essential physical information for accurate simulation. Third, we incorporated the direction parameters of the sun and the satellite, which addressed the dependence of VIS images on sunlight directions and enabled a much larger training set from continuous daytime data. The model was trained and validated using data from the advanced Himawari imager in the daytime, achieving statistical results of SSIM = 0.923 and root mean square error = 0.0299, which significantly surpasses existing models. We also performed a cross-satellite nighttime model validation using the day/night band of the visible/infrared imager radiometer suite, which yields outstanding results compared to existing models. Our model is operationally applied to generate accurate VIS imagery with arbitrary virtual sunlight directions, significantly contributing to the nighttime monitoring of various meteorological phenomena.https://ieeexplore.ieee.org/document/10988561/Advanced Himawari imager (AHI)cloudsconditional generative adversarial network (CGAN)deep learningnighttimetropical cyclone (TC)
spellingShingle Jinghuai Yao
Puyuan Du
Yucheng Zhao
Yubo Wang
Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Advanced Himawari imager (AHI)
clouds
conditional generative adversarial network (CGAN)
deep learning
nighttime
tropical cyclone (TC)
title Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks
title_full Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks
title_fullStr Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks
title_full_unstemmed Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks
title_short Simulating Nighttime Visible Satellite Imagery of Tropical Cyclones Using Conditional Generative Adversarial Networks
title_sort simulating nighttime visible satellite imagery of tropical cyclones using conditional generative adversarial networks
topic Advanced Himawari imager (AHI)
clouds
conditional generative adversarial network (CGAN)
deep learning
nighttime
tropical cyclone (TC)
url https://ieeexplore.ieee.org/document/10988561/
work_keys_str_mv AT jinghuaiyao simulatingnighttimevisiblesatelliteimageryoftropicalcyclonesusingconditionalgenerativeadversarialnetworks
AT puyuandu simulatingnighttimevisiblesatelliteimageryoftropicalcyclonesusingconditionalgenerativeadversarialnetworks
AT yuchengzhao simulatingnighttimevisiblesatelliteimageryoftropicalcyclonesusingconditionalgenerativeadversarialnetworks
AT yubowang simulatingnighttimevisiblesatelliteimageryoftropicalcyclonesusingconditionalgenerativeadversarialnetworks