Robust Haze and Thin Cloud Removal via Conditional Variational Autoencoders

Existing methods for remote-sensing image dehazing and thin cloud removal treat this image restoration task as a clear pixel estimation problem, yielding a single prediction result through a deterministic pipeline. However, image restoration is a highly ill-posed problem, as the sharp pixel value co...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on geoscience and remote sensing Jg. 62; S. 1 - 16
Hauptverfasser: Ding, Haidong, Xie, Fengying, Qiu, Linwei, Zhang, Xiaozhe, Shi, Zhenwei
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2024
Schlagworte:
Algorithms
Benchmarks
Haze
Ill posed problems
Image degradation
Image restoration
Pixels
Remote sensing
Removal
Restoration
ISSN:0196-2892, 1558-0644
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Existing methods for remote-sensing image dehazing and thin cloud removal treat this image restoration task as a clear pixel estimation problem, yielding a single prediction result through a deterministic pipeline. However, image restoration is a highly ill-posed problem, as the sharp pixel value corresponding to the input cannot be uniquely determined solely from the degraded image. In this article, we present a novel algorithm for haze and thin cloud removal using conditional variational autoencoders (CVAEs) to generate multiple realistic restored images for each input. By sampling from the latent space to capture the pixel diversity, the proposed method mitigates the limitations arising from inaccuracies in a single estimation. In this uncertainty pipeline, we can generate a more accurate restored image based on these multiple predictions. Furthermore, we have developed a dynamic fusion network (DFN) for combining multiple plausible outcomes to obtain a more accurate result. DFN dynamically predicts the kernels used for restored result generation conditioned on inputs, improving haze and thin cloud thanks to its adaptive nature. Quantitative and qualitative experiments demonstrate that the proposed method outperforms existing state-of-the-art techniques by a significant margin on dehazing and thin cloud removal benchmarks.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0196-2892
1558-0644
DOI:10.1109/TGRS.2024.3349779