Abstract
We analyzed large-scale geological hazards in the case of the Northern Pakistan Earthquake, the Middle Java Earthquake and the Leyte landslide, using optical high-resolution satellite imagery such as QuickBird, IKONOS, ALOS PRISM and SPOT5. The characteristics of satellite images were compared to those of disasters on the interpretation cards. It is possible to recognize large collapses using 2.5 m-resolution satellite images; however 1 m-resolution satellite images are necessary for small-scale collapses and cracks. For the recognition of building damages in the urban area, 1 m-resolution stereo images are necessary; however it is impossible by 1 m-resolution mono image. For surface earthquake faults, it is useful to combine the recognition of building heavy damaged area using 1 m-resolution satellite image, and the identification of cliff landforms using 2.5 m-resolution stereo images. In order to understand hazard conditions, it is important for satellite imagery to be not only of high resolution but also stereoscopic.
We proposed a method for recognition of large-scale geological hazards using optical high-resolution satellite imagery.