Comparative Study of Water Body Extraction Methods for Poyang Lake Based on Sentinel-1 SAR Data
Remote sensing technique plays a crucial role in water body planning and monitoring.This study focused on water body extraction using Sentinel-1 SAR data.Firstly,the backscattering characteristics of water bodies and their surrounding land features were analyzed.Secondly,in conjunction with texture and terrain attributes,the effectiveness of threshold,classifier,and object-oriented methods for water body extraction were compared.Subsequently,to explore the extraction of typical water environ-ments in Poyang Lake,application strategies for various methods were presented.The results indicate that the fusion of multiple features reduces ambiguity in water body information.The RF algorithm demonstrates precise extraction of small water bodies,achieving an overall accuracy of 97.8%.Threshold methods like SDWI and Otsu exhibit commendable recognition capabilities for shallow water areas with vegetation.Notably,ships on the water surface impact the completeness of water body extraction.This re-search discusses the applicability of Sentinel-1 SAR data for water body extraction,aiming to provide crucial technical support of monitoring water resources in Poyang Lake.