KAIST AI reconstructs cloud-covered ground in satellite images

A KAIST research team has developed an AI model that restores cloud-covered areas in satellite images and converts radar imagery into photo-like views. Trained on diverse terrain and cities, the model could expand satellite data use in disaster response and environmental monitoring.

• Model restores ground hidden by clouds in satellite images
• Radar images are converted into photo-like pictures
• KAIST expects use in disaster response and environmental monitoring