Landslide Monitoring from the Perspective of Surveillance Cameras with Change Detection Technology

Le résumé fourni par la source

Landslide is one of the most dangerous natural disasters, which cause significant economic and even human life loss every year. To reduce the loss caused by landslide, it is important to monitor the occurrence of landslide and give an alarm in time. At present, there are two mainstream technical routes for landslide monitoring, namely with IoT sensors and with satellite imagery. However, IoT sensors cannot monitor large area of land, while satellite image cannot alarm in time. In this paper, we present monitoring landslide with surveillance cameras, which are able to cover large area of land, and to send alarm promptly. To detect landslide on surveillance camera image, we use deep learning based change detection technology, which can tell newly occurred landslide from historical landslide scar. Furthermore, AIGC technology is used to build up a dataset to train the landslide change detection model.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.