CLOUD AND FOG REMOVAL FROM SATELLITE IMAGES USING GENERATIVE ADVERSARIAL NETWORKS (GANs)
Résumé
Advanced Earth observation satellites produce a constant stream of data. These datasets have wide applications in ecosystem research such as vegetation health monitoring, water quality, disaster monitoring, and a range of other areas. Among the earth observation satellites, multispectral remote sensing data have been widely used for the last several decades and continue to be at the center of earth observation datasets for the scientific communities. Multispectral remote sensing data suffer from environmental effects due to cloud and fog and hence a major chunk of it becomes unusable. Therefore, being able to use advanced AI/ML techniques to remove these environmental impurities from these data to produce uninterrupted datastream can be quite valuable. Here we demonstrate the use of Generative Adversarial Network (GAN) to remove fog and cloud from Landsat multispectral data to regenerate and fill in the gaps for uninterrupted satellite data.
Domaines
Intelligence artificielle [cs.AI]Origine | Fichiers produits par l'(les) auteur(s) |
---|