Vision Transformer-Based Time-Series Image Reconstruction for Cloud-Filling Applications

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Li, Lujun, Wang, Yiqun, State, Radu
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866910103795400704
author Li, Lujun
Wang, Yiqun
State, Radu
author_facet Li, Lujun
Wang, Yiqun
State, Radu
contents Cloud cover in multispectral imagery (MSI) poses significant challenges for early season crop mapping, as it leads to missing or corrupted spectral information. Synthetic aperture radar (SAR) data, which is not affected by cloud interference, offers a complementary solution, but lack sufficient spectral detail for precise crop mapping. To address this, we propose a novel framework, Time-series MSI Image Reconstruction using Vision Transformer (ViT), to reconstruct MSI data in cloud-covered regions by leveraging the temporal coherence of MSI and the complementary information from SAR from the attention mechanism. Comprehensive experiments, using rigorous reconstruction evaluation metrics, demonstrate that Time-series ViT framework significantly outperforms baselines that use non-time-series MSI and SAR or time-series MSI without SAR, effectively enhancing MSI image reconstruction in cloud-covered regions.
format Preprint
id arxiv_https___arxiv_org_abs_2506_19591
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Vision Transformer-Based Time-Series Image Reconstruction for Cloud-Filling Applications
Li, Lujun
Wang, Yiqun
State, Radu
Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Image and Video Processing
Cloud cover in multispectral imagery (MSI) poses significant challenges for early season crop mapping, as it leads to missing or corrupted spectral information. Synthetic aperture radar (SAR) data, which is not affected by cloud interference, offers a complementary solution, but lack sufficient spectral detail for precise crop mapping. To address this, we propose a novel framework, Time-series MSI Image Reconstruction using Vision Transformer (ViT), to reconstruct MSI data in cloud-covered regions by leveraging the temporal coherence of MSI and the complementary information from SAR from the attention mechanism. Comprehensive experiments, using rigorous reconstruction evaluation metrics, demonstrate that Time-series ViT framework significantly outperforms baselines that use non-time-series MSI and SAR or time-series MSI without SAR, effectively enhancing MSI image reconstruction in cloud-covered regions.
title Vision Transformer-Based Time-Series Image Reconstruction for Cloud-Filling Applications
topic Computer Vision and Pattern Recognition
Artificial Intelligence
Machine Learning
Image and Video Processing
url https://arxiv.org/abs/2506.19591