logo
CanyamAI summaries for academic research
language
MobileMobile
  • Phone FeatureWeChat Mini Program
EnglishEnglish
noticeNotice
BlogBlog
delete
arrow
Return

Physics-informed machine learning for cloud detection

delete2026-09-26
delete0
delete
OA
AI
S
Shi Qiu *
Z
Zhe Zhu *
X
Xiucheng Yang
J
Junchang Ju
Q
Qiang Zhou
C
C. S. R. Neigh
DOI：10.1016/j.rse.2026.115672delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• Fmask 5: Physics-informed ML for Landsat/Sentinel-2 cloud detection. • Physical rules provide dynamic, localized training for ML models. • Strategic LightGBM and UNet combo leverages pixel and CNN strengths. • Fmask 5 outperforms both physical-rule-only and ML-only cloud detection. • High accuracy, CPU-efficiency; suitable for operational production.
Keywords:
Cloud detection
Landsat
Sentinel-2
Fmask
PIML
LightGBM
UNet

Journal

Remote Sensing of Environment cover
Remote Sensing of Environment
IF:
11.4
Papers:
1.1W
Citations:
9.4W

Organization

U
University of Connecticut
Scholars:
321
Papers: 138
Citations: 0
N
nasa goddard space flight center
Scholars:
70
Papers: 37
Citations: 0
Cited Papers

Cited Papers

No cited papers available