Pemetaan Habitat Dasar Perairan Dangkal Menggunakan Citra Satelit WorldView-2 dengan Skema Klasifikasi Supervisi dan Koreksi Kolom Air di Perairan Pulau Tunda.

Abstract

Remote sensing satellite is an ideal platform to get geospatial information of shallow sea waters. The accuracy of the image is a fundamental problem in benthic habitat mapping shallow waters. This research aims to mapping benthic habitat from WorldView-2 imagery with scheme classification in Tunda Island. The method used was image classification guided by the composite channel and image enhancement with depth invariant index algorithms. Field survey was conducted to observe directly the existing objects in the image. The bentihc habitats were found as many as 10 habitat classes then reduced to 6 habitat classes. Image classification shows the basic habitat of rubble and combination of coral and rubble almost dominated the whole area of observation. The accuracy test results of supervised classification with depth invariant index showed the greater accuracy values than using other with the value reached 74.42% in 6 habitat classes scheme.