Disaster detection and damage estimation using satellite imagery and land-use information

Abstract

In the paper, the damage detection and land-use analysis for a major typhoon, called AERE, will be addressed. The damage detection is based on regular monitoring of the land-cover changes in the watershed. A mechanism for periodical data acquisition, ortho-images rectification and land-cover change detection is performed for the reservoir management. The mechanism can also assure a fast response for disaster that happened in the watershed. In the mechanism, an intensive data acquisition in the watershed using the SPOT images with a frequency of three images per week has been achieved. One dataset of every two weeks is selected. Images with less than 50% cloud cover are chosen for ortho-rectification and further for land-cover change detection. When the changes of vegetation indices for multi-temporal images exceed a ceiling, the location is subject to in-situ inspection to prevent deteriorations. Illegal human activities and land-cover changes were successfully supervised and controlled. If a disaster happens and causes damage, the integration of satellite images and relevant geospatial data, such as land-use maps and terrain models will then provide a good solution for damage estimation. The spatial relationship between the damage and the land-use type can thus be analyzed.