Coral reefs are among the most valuable marine ecosystems in the world, yet they are increasingly threatened by human activities and climate change. To protect these ecosystems, accurate and efficient monitoring methods are essential. A study conducted in the Spermonde Archipelago, South Sulawesi, Indonesia, demonstrates how satellite imagery can be used to assess coral reef health effectively.
The research compares two widely used unsupervised image classification methods—IsoData and K-Means—combined with the Lyzenga algorithm to analyze shallow-water habitats using SPOT-6 and Sentinel-2A satellite images. These methods allow coral reef conditions to be mapped without requiring extensive prior field data, making them practical for large and remote coastal areas.
The results show that both IsoData and K-Means achieved an overall classification accuracy of 62.5%, meeting commonly accepted standards for shallow-water habitat mapping. However, the two methods revealed different strengths. IsoData was more effective in identifying live coral and algae, while K-Means performed better in detecting rubble and seagrass areas. This highlights the importance of choosing the right classification approach depending on the monitoring objective.
The study also found that many damaged reef areas were associated with human activities, such as destructive fishing practices, as well as natural stressors like coral bleaching. By combining satellite data with field observations, the research provides valuable spatial information that can support coral reef management, conservation planning, and long-term monitoring.
This work demonstrates the strong potential of remote sensing and geospatial analysis as cost-effective tools for coral reef assessment in Indonesia and other tropical regions. The findings offer practical guidance for researchers, coastal managers, and policymakers seeking reliable methods to monitor reef health and support sustainable marine management.
