TEC Map Completion Using DCGAN and Poisson Blending

Abstract Because of the limited coverage of global navigation satellite system (GNSS) receivers, total electron content (TEC) maps are not complete. The processing to obtain complete TEC maps is time consuming and needs the collaboration of five international GNSS service (IGS) centers to consolidat...

Full description

Saved in:
Bibliographic Details
Main Authors: Yang Pan, Mingwu Jin, Shunrong Zhang, Yue Deng
Format: Article
Language:English
Published: Wiley 2020-05-01
Series:Space Weather
Subjects:
Online Access:https://doi.org/10.1029/2019SW002390
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1841536403883687936
author Yang Pan
Mingwu Jin
Shunrong Zhang
Yue Deng
author_facet Yang Pan
Mingwu Jin
Shunrong Zhang
Yue Deng
author_sort Yang Pan
collection DOAJ
description Abstract Because of the limited coverage of global navigation satellite system (GNSS) receivers, total electron content (TEC) maps are not complete. The processing to obtain complete TEC maps is time consuming and needs the collaboration of five international GNSS service (IGS) centers to consolidate final completed IGS TEC maps. The advance of deep learning offers powerful tools to perform certain tasks in data science, such as image completion (or inpainting). Among them, deep convolutional generative adversarial network (DCGAN) is capable of learning the properties of the objects and recovering missing data effectively. With years of IGS TEC maps for training, the combination of DCGAN and Poisson blending (DCGAN‐PB) is able to effectively learn the completion process of IGS TEC maps. Both random and more realistic masks are used to test the performance of DCGAN‐PB. The results with random masks (15–40% missing data) show that DCGAN‐PB can achieve better TEC map completion than DCGAN alone, and more training data can significantly improve its generalization. For the cross‐validation experiment using the realistic mask from Massachusetts Institute of Technology (MIT)‐TEC data (~52% missing data), DCGAN‐PB achieves the average root mean squared error about three absolute TEC units (TECu) for high solar activity years and less than two TECu for low solar activity years, which is about 50% reduction of means and more than 50% reduction on standard deviations compared to two conventional single‐image inpainting methods. The DCGAN‐PB model can lead to an efficient automatic completion tool for TEC maps by minimizing the manual work.
format Article
id doaj-art-59b19ac09724400193b08d83f96c7db3
institution Kabale University
issn 1542-7390
language English
publishDate 2020-05-01
publisher Wiley
record_format Article
series Space Weather
spelling doaj-art-59b19ac09724400193b08d83f96c7db32025-01-14T16:27:35ZengWileySpace Weather1542-73902020-05-01185n/an/a10.1029/2019SW002390TEC Map Completion Using DCGAN and Poisson BlendingYang Pan0Mingwu Jin1Shunrong Zhang2Yue Deng3The Department of Physics University of Texas at Arlington Arlington TX USAThe Department of Physics University of Texas at Arlington Arlington TX USAHaystack Observatory Massachusetts Institute of Technology Westford MA USAThe Department of Physics University of Texas at Arlington Arlington TX USAAbstract Because of the limited coverage of global navigation satellite system (GNSS) receivers, total electron content (TEC) maps are not complete. The processing to obtain complete TEC maps is time consuming and needs the collaboration of five international GNSS service (IGS) centers to consolidate final completed IGS TEC maps. The advance of deep learning offers powerful tools to perform certain tasks in data science, such as image completion (or inpainting). Among them, deep convolutional generative adversarial network (DCGAN) is capable of learning the properties of the objects and recovering missing data effectively. With years of IGS TEC maps for training, the combination of DCGAN and Poisson blending (DCGAN‐PB) is able to effectively learn the completion process of IGS TEC maps. Both random and more realistic masks are used to test the performance of DCGAN‐PB. The results with random masks (15–40% missing data) show that DCGAN‐PB can achieve better TEC map completion than DCGAN alone, and more training data can significantly improve its generalization. For the cross‐validation experiment using the realistic mask from Massachusetts Institute of Technology (MIT)‐TEC data (~52% missing data), DCGAN‐PB achieves the average root mean squared error about three absolute TEC units (TECu) for high solar activity years and less than two TECu for low solar activity years, which is about 50% reduction of means and more than 50% reduction on standard deviations compared to two conventional single‐image inpainting methods. The DCGAN‐PB model can lead to an efficient automatic completion tool for TEC maps by minimizing the manual work.https://doi.org/10.1029/2019SW002390TEC mapsDCGANdeep learningmap completionPoisson blendingionosphere
spellingShingle Yang Pan
Mingwu Jin
Shunrong Zhang
Yue Deng
TEC Map Completion Using DCGAN and Poisson Blending
Space Weather
TEC maps
DCGAN
deep learning
map completion
Poisson blending
ionosphere
title TEC Map Completion Using DCGAN and Poisson Blending
title_full TEC Map Completion Using DCGAN and Poisson Blending
title_fullStr TEC Map Completion Using DCGAN and Poisson Blending
title_full_unstemmed TEC Map Completion Using DCGAN and Poisson Blending
title_short TEC Map Completion Using DCGAN and Poisson Blending
title_sort tec map completion using dcgan and poisson blending
topic TEC maps
DCGAN
deep learning
map completion
Poisson blending
ionosphere
url https://doi.org/10.1029/2019SW002390
work_keys_str_mv AT yangpan tecmapcompletionusingdcganandpoissonblending
AT mingwujin tecmapcompletionusingdcganandpoissonblending
AT shunrongzhang tecmapcompletionusingdcganandpoissonblending
AT yuedeng tecmapcompletionusingdcganandpoissonblending