Earthquake-induced geological hazards detection under hierarchical stripping classification framework in the Beichuan area

Abstract

The Wenchuan earthquake-induced geological hazards, such as rock falls, landslides, and debris flows, caused as much damage as the earthquake itself. In the complicated environment of the disaster area, seldom can a single remotely sensed imagery and a single classification method meet the needs of the interpretation task. In this paper, taking the Beichuan area as an example, the hierarchical stripping classification (HSC) framework, supported by human-computer interactive interpretation, was applied to tackle the problems caused by the earthquake, as well as the local complicated landform and weather environments. Different classification tools were used to detect specific objects under the HSC framework. Multi-source and multi-temporal data from satellite remote sensors (mainly CBERS-02B charge-coupled device and HR images), unmanned aerial vehicles (UAV), digital elevation model, and other auxiliary sources were utilized to detect the distribution of geological hazards. The classification results were compared with that of an overall supervised classification. The results show that the HSC framework and the related methods are effective and work well in the study area.