Case Study: University of Salzburg, Austria achieves rapid, accurate landslide mapping with Trimble eCognition

How eCognition is helping researchers gain firm footing in the unpredictable landscape of identifying, mapping, and potentially, predicting landslides

Trimble’s eCognition software is being used by researchers (including the University of Salzburg and New Zealand’s Landcare Research) to tackle the difficult problem of identifying, mapping and ultimately predicting landslides. Landslides are varied, often hard to distinguish from manmade features, and traditionally mapped by slow, subjective manual interpretation of aerial and satellite images, creating a need for faster, more consistent tools and comprehensive inventories.

Using an object‑based image analysis (OBIA) approach in eCognition, teams developed rule sets to semi‑automate landslide detection across time-series orthophotos and to integrate optical imagery with InSAR and DEM data. The automated method matched manual mapping accuracy but reduced processing from weeks to hours, improved consistency and repeatability, flagged unstable areas, and—by combining visible and underground movement data—enabled more complete inventories and the potential for predictive modelling.

University of Salzburg, Austria

Daniel Hölbling

Research Scientist, Department of Geoinformatics