Research on Information Extraction of Collapsed Houses by Floods

Flood damage assessment of houses collapse is very important. At present, disaster and loss information are obtained mainly on the report from local government. However, some information may be delayed or have man-made errors. In this paper, remote sensing (RS) data are used to carry out research on information extraction of flood damaged houses. Different sources (optical, radar) and different resolutions (high, medium, low) of RS data are evaluated of their usage in the work of collapsed houses extraction. The suitability of various data is analyzed and combinations of possible sources are recommended. Results showing that multi-spectrum RS data with resolution better than 3 meters may have better result compared with other optical sources. Radar data have advantages in the extraction during the flood period. Research results may provide decision-making support to the government for rapid information extraction and assessment of house losses caused by floods.