Object based classification of benthic habitat using Sentinel 2 imagery by applying with support vector machine and random forest algorithms in shallow waters of Kepulauan Seribu, Indonesia

Abstract. Hartoni, Siregar VP, Wouthuyzen S, Agus SB. 2021. Object based classification of benthic habitat using Sentinel 2 imagery by applying with support vector machine and random forest algorithms in shallow waters Kepulauan Seribu. Biodiversitas 23: 514-520. Benthic habitats have very high complexity are home to many types aquatic organisms. various functions, including for flora fauna, sediment traps, nursery areas, foraging areas fauna that susceptible damage due human activities or natural factors. Therefore, more accurate spatial information is needed. The purpose this study was examine the ability object-based techniques mapping 2A imagery. two used (SVM) (RF). input image layer (IIL) color band (Band 432). results showed SVM RF could classify eight classes habitats. overall accuracy (OA) algorithm 65%, while 67%, kappa values 0.59 0.60, respectively. significant test applied images has a Z value of-0.41. These indicate between not significantly different.