Details

Autor(en) / Beteiligte

• Rewehel, Ekram Mokhtar
• Li, Jianqiang
• Keshk, Hatem M.

Titel

RETRACTED ARTICLE: Sentinel-2 Cloud Mask Classification Using Deep Learning Method

Ist Teil von

• International journal of aeronautical and space sciences, 2022-07, Vol.23 (3), p.622-635

Ort / Verlag

Seoul: The Korean Society for Aeronautical & Space Sciences (KSAS)

Erscheinungsjahr

2022

Link zum Volltext

Quelle

Alma/SFX Local Collection

Beschreibungen/Notizen

• The Copernicus Sentinel-2 mission operated by the European Space Agency (ESA) provides comprehensive and continuous multi-spectral observations of all Earth’s land surface since mid-2015. Clouds and cloud shadows significantly decrease the usability of optical satellite data especially in agricultural applications; therefore, an accurate and reliable cloud mask is mandatory for effective Earth Observation (EO) optical data exploitation. During the last years, image segmentation techniques have developed rapidly with the exploitation of neural networks capabilities. In this perspective, the KappaMask processor using U-Net architecture was developed with the ability to generate a classification mask over northern latitudes into the following classes: clear, cloud shadow, semi-transparent cloud (thin clouds), and cloud, invalid. For training, a Sentinel-2 dataset covering Northern European terrestrial area was labeled. KappaMask provides a 10 m classification mask for Sentinel-2 Level- 2A (L2A) and Level-1C (L1C) products. The total dice coefficient on the test dataset, which was not seen by the model at any stage, was 80% for KappaMask L2A and 76% for KappaMask L1C for 27 clear, cloud shadow, semi-transparent, cloud classes. Comparison with rule-based cloud mask 28 methods was then performed on the same test dataset, where Sen2Cor reached 59% dice coefficient 29 for clear, cloud shadow, semi-transparent, cloud classes, Fmask 61% for clear, cloud shadow, cloud 30 classes, and Maja 51% for clear and cloud classes. The closest machine learning open-source cloud classification mask, S2cloudless, has 63% dice coefficient providing only cloud and clear classes, while KappaMask L2A with a more complex classification schema outperformed S2cloudless by 17%.