Landslide mapping using object-based image analysis and open source tools

Abstract:

Landslides are pervasive issue across the world, causing millions of dollars’ worth of damage to infrastructure, severe economic losses, and thousands of deaths annually. A robust and complete landslide inventory is often the first requirement for quantifying landslide hazard and risk. Currently, the most used method to map landslides is manual mapping, a technique that is limiting in space and time. Recent availability of very high-resolution optical imagery and advancement in image processing technologies have significantly improved our ability to map landslides. In recent years object-based image analysis (OBIA) has been gaining in popularity for landslide mapping due to its ability to incorporate spectral, textural, morphological and topographical properties. We created a Semi-Automatic Landslide Detection (SALaD) system in a Python environment utilizing OBIA and machine learning. It uses various open source Python packages and modules and can be configured in both Linux and Windows environments. The SALaD system was tested by creating multitemporal landslide inventories over major transportation networks in Nepal as well as event-based landslide mapping all over the world. Performance and challenges of the SALaD system will be highlighted and discussed.