Abstract:

This paper presents a method for change detection using multitemporal synthetic aperture radar images based on statistical similarity measure in order to determine inundation regions. From Ground Range Detected (GRD) products of SAR Sentinel-1 satellite acquired before, during and after the the flood caused by typhoon Damrey in Hue city and a part of Huong Tra district, Thua Thien Hue province, changes on the ground were detected using Kullback-Leibler distance between two Log-normal distributions. Inundation detection results were evaluated by validation samples derived from optical 3 m resolution PlanetScope image, and the obtained accuracy is 87,3%. Change information extracted from multitemporal SAR images then could be exploited for different tasks, such as: inundation region mapping, statistical analysis of affected areas, etc