Photogrammetric monitoring concept for remote landslide endangered areas using multi-temporal aerial imagery

Natural disasters like landslides or floods can be significant dangers to life, nature and infrastructure in the affected areas. Especially in regions with permanent potential risks of such weather-induced phenomena, an effective monitoring system can reduce the risk of major damages. In the case of Punta Arenas (Southern Chile) landslides irregularly block and dam the nearby river Las Minas and thus cause catastrophic floods of the Chilean city. To prevent such landslide triggered hazards, the paper focuses on photogrammetric landslide detection and analysis using multi-temporal aerial images to reliably detect major geomorphological changes. As example, datasets from 2014 and 2015 of the river Las Minas have been compared in order to detect volume changes. For this purpose, multi-temporal 3D point clouds were generated and georeferenced. The registration of the point clouds was conducted applying ICP techniques to compensate potential point cloud deformations.