Rakit Island in Saleh Bay, West Nusa Tenggara, possesses shallow marine ecosystems that are ecologically important but remain under-studied. This study aimed to map shallow marine benthic habitats in the waters of Rakit Island using Sentinel-2A satellite imagery. An object- based image analysis (OBIA) approach combined with a Support Vector Machine (SVM) classification algorithm was applied. The methodological workflow included atmospheric correction, water column correction, multiresolution segmentation, a two-level classification process, and accuracy assessment using field validation data. The classification results identified seven benthic habitat classes, namely rocks, sand, muddy sand, seaweed, debris, live coral, and dead coral with algae. The overall classification accuracy reached 69.01%, with a kappa coefficient of 0.63, indicating a good level of agreement between the classification results and field observations. The main limitations were spectral similarity among habitat classes and the influence of water turbidity, particularly affecting seaweed detection in deeper waters. Overall, the results demonstrate that the OBIA–SVM approach is effective for mapping shallow marine habitats using medium-resolution Sentinel-2A imagery.
Maulana et al. (Wed,) studied this question.