Flood damage monitoring on land use using radar and optical satellite image processing

This paper aims to estimate flood damage on land use areas through a three-step process: identification of flood areas using Sentinel-1 images, mapping of land use using Sentinel-2 images, and estimation of damages in each land use class. The case study is located in the north of Ahvaz City in Khuzestan Province. To generate the flood extent map, the water index from Sentinel-1 images before and after the flood is obtained by fusing VV and VH polarizations using the discrete wavelet transform method. Then, the water index image of each time is segmented into superpixels using Simple Linear Iterative Clustering (SLIC). Afterward, the obtained superpixels are classified using K-means clustering into water bodies and backgrounds. The results obtained by classifying Sentinel-2 images indicate that agricultural lands are the most prone to damage. In total, 27% of the areas were destroyed, and 26% of these damages belong to agricultural lands. Moreover, more than 50% of agricultural lands are covered by type 2 and 3 agricultural fields.