Semiautomated object-based classification of rain-induced landslides with VHR multispectral images on Madeira Island

Abstract

A method for semiautomated landslide detection and mapping, with the ability to separate source and run-out areas, is presented in this paper. It combines object-based image analysis and a support vector machine classifier and is tested using a GeoEye-1 multispectral image, sensed 3 days after a major damaging landslide event that occurred on Madeira Island (20 February 2010), and a pre-event lidar digital terrain model. The testing is developed in a 15 km2 wide study area, where 95% of the number of landslides scars are detected by this supervised approach. The classifier presents a good performance in the delineation of the overall landslide area, with commission errors below 26% and omission errors below 24%. In addition, fair results are achieved in the separation of the source from the run-out landslide areas, although in less illuminated slopes this discrimination is less effective than in sunnier, east-facing slopes.