An Ensemble of Classifiers Approach to Coral Distribution Mapping

Accurate, up-to-date and accessible information on the state of coral reef ecosystem is necessary for informed and effective management of these important marine resources. However, environments containing these habitats are challenging to map due to their remoteness, extent and costs of monitoring. In this research, the capabilities of satellite remote sensing techniques combined with in situ data were assessed to generate coral habitat map of Lang Tengah Island. Several classification techniques were utilized in identifying coral distribution. This study aims at utilizing a new approached to increase accuracy, mainly using an ensemble of classifier. The core principle of the ensemble methodology is to weigh several individual classifiers, and combine them in order to reach a classification that is better than the one obtained by only one classifier. This ensembles offer promise in boosting the overall classification accuracy. Initially 5 classifiers were used to classify the study area mainly, Parallelepiped, Minimum distance, Maximum likelihood, Fisher and K-Nearest Neighbor. Results from the classification shows that each method produced different accuracy based on the bottom type. The K-Nearest Neighbor method produced the most accurate results for dense coral class (71.43 %) while for sand class Maximum likelihood was the most accurate (70.37%). Using these results an ensemble of classification approach was applied and results show that accuracy was the highest with an overall accuracy of 73.02 % in comparison to Parallelepiped (52.38), Minimum distance (50.79), Maximum likelihood (60.37), Fisher (31.75) and KNearest Neighbor (50.79) Keyword: Ensemble of Classifier, Hard Classification, Coral Reef Mapping.