GRSG Conference 2022: Orbit to Outcrop

Poster Title: Towards multitemporal landslide detection through cloud computing

Author: Muhammad Aufaristama


Change detection from two pairs of images (pre and post-event) is commonly used for landslide classification mapping. The question is if we have many images of pre and post-landslide events, Can we consider the time component to detect and map landslides?

This research aims to explore and take advantage of the Google Earth Engine (GEE) cloud computing tools to process large time series datasets for automated landslide detection. In summary, this research will demonstrate the power of the GEE platform in handling large time series datasets and developing automated mapping web applications based on expert knowledge and existing techniques that are commonly used for landslide mapping.

Performance and limitations of the tool, such as its accuracies, combined with other techniques, aiming to develop the user-friendly application and extend the temporal range of data analyses, are discussed.