FLOODED AREAS EXTRACTION DUE TO THE 2011 THAILAND FLOOD USING RADARSAT-2 AND THAICHOTE IMAGERY DATA

摘要

This paper examines an extraction method of widespread flooded areas occurred in the Chao Phraya River basin, central Thailand, in the rainy season of 2011. RADARSAT-2 imagery data have been mainly used to extract affected areas, while THAICHOTE imagery data have been used as optical supporting data for the Thai Government. In this study, the same data were used in a somewhat different method with more deeply in detail. ScanSAR Narrow-mode imagery with cross-polarization of RADARSAT-2 was introduced to improve the accuracy and get more information on the ground surface. The SAR intensity images, which can be acquired also in the nighttime or under bad weather conditions, were found to be the most effective because the smoothness of water surface always shows low backscatter values. In the same way, the NDVI values calculated from the THAICHOTE images could also recognize flooded areas form open space under a clear sky condition. However, both of these sensors could not discriminate flooded urban areas easily because of the limitation of their spatial resolutions. Backscatter values still kept high although buildings were surrounded by water. The extracted results were validated by a high-resolution optical satellite image, water height data from gaging stations and a digital surface model (DEM) from LiDAR.