A new research project at the University of Twente will develop a novel way to identify and monitor landslides, even when cloud cover obscures traditional satellite images. This breakthrough could significantly improve landslide warning systems and save lives around the world.
Landslides caused by earthquakes, heavy rain, and human activities threaten lives, infrastructure, and the environment. They cause an average of 4,500 deaths and 20 billion USD in damage annually. In February 2023, earthquakes in Turkey and Syria showed the devastation landslides can cause.
Near-real-time landslide detection
Current landslide mapping methods rely on cloud-free optical satellite images, which are often scarce in mountainous regions. This research will use an AI-based smart tracking technique to develop a prototype open-access landslide catalogue. The new methodology will enable near-real-time landslide detection and monitoring, allowing early warning systems to be activated to protect communities at risk. This could save lives and reduce damage from landslides.
"This innovative research project has the potential to advance landslide monitoring and prediction significantly," said Serkan Girgin, lead researcher of the project. "By developing a methodology leveraging big data processing capabilities and machine learning methods to identify and track landslides independently of cloud cover, we can provide valuable insights into landslide activity globally. Furthermore, by creating an up-to-date landslide registry by operationalizing the methodology we will facilitate mitigation of risks and save lives."
More information
Serkan Girgin is associate professor at the Department of Geo-information Processing (GIP; Faculty of ITC). He is also head of the Centre of Expertise in Big Geodata Science (CRIB).
The funding for this research project, entitled ‘Landslide Hunter: the first fully automated AI-based platform to map and monitor landslides remotely’, has been awarded by the Dutch Research Council (NWO) as part of the Open Competition – XS Round 4, where out of 190 eligible applications only 28 proposals were funded. The proposal ranked 2nd in cluster Physical and Exact Sciences Group B. The project is expected to start in early 2024 and run for a year.
