We present a method for cloud-removal from satellite images using axial transformer networks. The method considers a set of multitemporal images in a given region of interest together with the corresponding cloud masks, and delivers a cloud-free image for a specific day of the year. We propose the combination of an encoder-decoder model employing axial attention layers for the estimation of the low-resolution cloud-free image, together with a fully parallel upsampler that reconstructs the image at full resolution. The method is compared with various baselines and state-of-the-art methods on two Sentinel-2 datasets, showing significant improvements across multiple standard metrics used for image quality assessment.
CLOUDTRAN: CLOUD REMOVAL FROM MULTITEMPORAL SATELLITE IMAGES USING AXIAL TRANSFORMER NETWORKS / Christopoulos, D.; Ntouskos, V.; Karantzalos, K.. - 43:2-2022(2022), pp. 1125-1132. (Intervento presentato al convegno 2022 24th ISPRS Congress on Imaging Today, Foreseeing Tomorrow, Commission II tenutosi a Nice, France) [10.5194/isprs-archives-XLIII-B2-2022-1125-2022].
CLOUDTRAN: CLOUD REMOVAL FROM MULTITEMPORAL SATELLITE IMAGES USING AXIAL TRANSFORMER NETWORKS
Ntouskos V.;
2022
Abstract
We present a method for cloud-removal from satellite images using axial transformer networks. The method considers a set of multitemporal images in a given region of interest together with the corresponding cloud masks, and delivers a cloud-free image for a specific day of the year. We propose the combination of an encoder-decoder model employing axial attention layers for the estimation of the low-resolution cloud-free image, together with a fully parallel upsampler that reconstructs the image at full resolution. The method is compared with various baselines and state-of-the-art methods on two Sentinel-2 datasets, showing significant improvements across multiple standard metrics used for image quality assessment.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
