Earth Sciences

Automatic detection of methane emissions in multispectral satellite imagery using a vision transformer

B. Rouet-leduc and C. Hulbert

This groundbreaking research by Bertrand Rouet-Leduc and Claudia Hulbert unveils a powerful deep learning approach utilizing Sentinel-2 multispectral satellite data to detect methane emissions. With a revolutionary Vision Transformer architecture, this method dramatically enhances detection abilities, identifying methane sources as small as 0.01 km². Discover how this innovative model outperforms existing techniques and proves effective across diverse environments.

00:00

Playback language: English

Index

Abstract

This paper presents a deep learning approach for detecting methane emissions using multispectral satellite data, specifically Sentinel-2. The method overcomes limitations of existing techniques by leveraging a Vision Transformer (ViT) based encoder-decoder architecture to achieve high spatial and temporal resolution detection with global coverage. The model demonstrates an order of magnitude improvement over state-of-the-art methods, detecting methane point sources down to plumes of 0.01 km², equivalent to 200-300 kg CH₄ h⁻¹ sources. Validation against airborne measurements confirms the model's effectiveness in diverse environments.

Publisher

Nature Communications

Published On

May 14, 2024

Authors

Bertrand Rouet-Leduc, Claudia Hulbert

Related Publications

Explore these studies to deepen your understanding of the subject.

Medicine and Health

Detection of cerebrospinal fluid biomarkers changes of Alzheimer's disease using a cognitive stress test in persons with subjective cognitive decline and mild cognitive impairment

M. Valles-salgado, M. J. Gil-moreno, et al.

Medicine and Health

Rapid and Sensitive Detection of Staphylococcus aureus in Blood Cultures Using a Nuclease-Activatable Fluorescent Probe

A. N. 1 and A. N. 2

Medicine and Health

Insights in the regulation of trimetylamine N-oxide production using a comparative biomimetic approach suggest a metabolic switch in hibernating bears

T. Ebert, J. Painer, et al.

Education

A qualitative study of pre-service teachers’ experienced benefits and concerns of using motivational interaction in practice after a training course

E. Renko, A. Koski-jännes, et al.

Listen, Learn & Level Up

Over 10,000 hours of research content in 25+ fields, available in 12+ languages.

No more digging through PDFs, just hit play and absorb the world's latest research in your language, on your time.

listen to research audio papers with researchbunny