KLASIFIKASI HABITAT PERAIRAN DANGKAL DARI CITRA MULTISPASIAL DI PERAIRAN PULAU KAPOTA DAN PULAU KOMPOONE, KEPULAUAN WAKATOBI

Shallow water habitat mapping is important to do because: (1) it can support the planning, management, and decision making of government spatial; (2) it can support and design a Marine Protected Area (MPA); (3) it can conduct a scientific research program to determine a knowledge about benthic ecosystem and seabed geology; (4) it can do seabed resource valuation, both biotic and abiotic, for economic and management goals. Nowadays, the standardization of thematic map details level in coastal ecosystem has not determined, especially in shallow water habitat based on coastal management needs in certain scale. The study aims to compare map accuracy level between SPOT 6, Sentinel 2A, and Landsat 8 classification results using support vector machine algorithm. The study site is in Wakatobi Island, including Kapota Island and Kompoone Island. The in-situ data took on July 2019. The 347 ground truth and transect images in the field analyzed using Coral Point Count with Excel Extension (CPCe). The classification scheme that was gotten is 8 habitat benthic classes, then conducted classification with classify them to be 6 and 5 classes. The result from SPOT 6 for 5 habitat classes has the highest overall accuracy. The differences between pixel (spatial resolution) and the amount of classification scheme influence accuracy results.