Bottom-type classification in coral reef area using hyperspectral bottom index imagery

This paper describes a bottom-types classification method for mapping shallow coral reef area using hyperspectral bottom index (BI) imagery. Monitoring of coral reef degradation, especially coral bleaching, is important because coral reefs offer important ecosystem services. However, it is difficult to monitor coral reefs using spaceborne multispectral sensors due to spectral confusion between coral, algae and seagrass. This study develops a bottom-type area ratio regression model using support vector machine (SVM) and hyperspectral BI imagery. BI imagery is effective in reducing wavelength dependency. The method applied in this study achieved overall accuracies that range from 0.6 to 0.8. In addition, the effectiveness of the applied method was confirmed with sample size reduction. For example, half of the sample size produced an overall accuracy of 0.6. This result shows that the applied method is effective for coral reef monitoring.