Accuracy assessment of relative and absolute water column correction methods for benthic habitat mapping in Parang Island

Abstract

The condition of benthic habitats in optically shallow sea waters becomes important information in the inventory and processing of coastal resources. Remote sensing is effective and efficient in mapping benthic habitats. This study aims to apply absolute and relative water column correction methods in order to map benthic habitats on Parang Island using PlanetScope image. The benthic habitat classification scheme used consists of coral reefs, seagrass, macroalgae, and substrate. We compared the accuracy of benthic habitat map based on absolute and relative water column correction methods. The classification methods used are the Maximum Likelihood (ML) algorithm and Support Vector Machine (SVM). The results showed that benthic habitat map with the highest accuracy was obtained by a combination of Lyzenga-ML at 61.63% followed by Purkis-SVM at 59.18%, Lyzenga-SVM at 41.90%, and Purkis-ML 16.87%. The results show that the Lyzenga water column correction method is the best choice in mapping benthic habitats.