AeroSat Vision Lab Pioneering AI and Satellite Innovation

Development of a Semi-Supervised Learning Model for Disaster Object Reconstruction in Partially Cloud-Covered Areas Using Edge Computing on a CanSAT Platform

Date: March 21, 2025

Development of a Semi-Supervised Learning Model for Disaster Object Reconstruction in Partially Cloud-Covered Areas Using Edge Computing on a CanSAT Platform

Tags: NSTC, 國科會計劃

This research aims to develop a semi-supervised learning model that reconstructs disaster-affected areas obscured by cloud cover. By integrating this model into edge computing devices onboard satellites, the project seeks to enhance real-time disaster detection and prediction capabilities.

This project explores semi-supervised learning for reconstructing disaster-affected regions covered by clouds.

  • AI-powered reconstruction of satellite images.
  • Integration with CanSAT, a small satellite platform.
  • Enhancing real-time disaster monitoring from space.

See more details on GitHub.