Improved Spectral Water Index Combined with Otsu Algorithm to Extract Muddy Coastline Data

Based on the spectral reflection characteristics analysis of muddy coastline in Jiangsu, an improved water index (IWI) combined with Otsu algorithm is proposed to extract coastlines from Landsat Operational Land Imager (OLI) images. The IWI-extracted results are compared those extracted by modified normalized difference (MNDWI), (NDWI), enhanced (EWI), revised different (RNDWI) and automated extraction (AWEI). show that IWI not affected tidal conditions or sand content water, can reduce “salt-and-pepper” phenomenon classification, accurately identify boundaries silty mudflats marine buildings high accuracy. It also significantly increase degree automation extraction. demonstrates accuracy over 84% data one-pixel tolerance, which twice as accurate other indices. for all types 81%. Therefore, reliable studies sea–land processes evolutions.