Flood Detection And Mapping From MultiTemporal Sentinel 1a Synthetic Aperture Radar Imagery Using Automated Flood Delineation Technique

Abstract

Precise and timely information about the flood extent mapping is critical to disaster management, humanitarian relief work and decision- making. There is a need for prominent flood detection algorithm to evaluate the flood extent and monitor the surface water changes using near-real-time satellite SAR imagery to produce timely and efficient results. newline newline newline newlineDuring the onset of the disaster, adequate number of image(s) might not be readily available despite the increasing availability of SAR acquisitions, especially for fine resolution imagery. Field data are rare and availability of cloud free optical imagery during the flood is difficult. To address this gap, this research presents an automated algorithm for delineating flood in areas when only a few SAR acquisitions are available. The proposed automated open surface water bodies extraction technique is both quick and simple, delineates water areas efficiently. The proposed algorithm combines Fuzzy C-means clustering (FCM) algorithm with binary thresholding (BT) and morphological operation (MO). This newline newline newline newline newline newline newlinealgorithm can be applied over large geographical areas and can compete with existing algorithms in terms of accuracy. newline newline newline newlineUsing multi-temporal Vertical transmit-Vertical receive (VV) polarized Synthetic Aperture Radar (SAR) imagery collected over few of the flood impacted districts of the state of Uttar Pradesh between 13th newlineAugust and 9th September 2017 by Sentinel-1A SAR Satellite, riverine newline newline newlineflood is analyzed and inundated areas are determined using the proposed flood delineation algorithm. Flood extent layers derived for the same study area from high resolution (HR) Horizontal transmit-Horizontal receive (HH) polarized SAR imagery for the dates 25th and 28th August newline2017 are used as reference flood data to validate the obtained results. newline newline newline newline newline newlineFor flood extent detection and mapping, pre-processing steps including sub-setting, image registration, and speckle noise suppression are performed for all the input SAR imagery. The processing stage is di