Elucidating the potential of unmanned aerial vehicle (UAV) and satellite imagery for the improvement of coral reef community mapping

Coral reefs are essential marine ecosystems that are becoming more and more endangered due to human activity, climate change, and coral bleaching. This emphasizes the importance of efficient monitoring and conservation measures. However, the low resolution of satellite data for fine-scale habitat detection and the restricted spatial coverage of unmanned aerial vehicles (UAVs) make mapping reefs difficult. In order to enhance coral reef classification in the coastal waters of the Universiti Malaysia Sabah (UMS) Outdoor Development Centre (ODEC) beach region, this study combines in-situ data, UAV footage, and PlanetScope satellite data. To evaluate the integration of UAV and satellite datasets, a five-step methodology was created. While resampled UAV datasets allowed for large-scale classification from PlanetScope imagery, resulting in an accuracy of 75.94%, combining in-situ and UAV data provided a coral classification map with 88.16% accuracy. Three separate reef zones – reef flat, reef crest, and fore reef – were distinguished on the resulting maps. Results show that large-scale satellite-based reef mapping can be improved by fine-scale UAV-derived datasets, providing a scalable way to monitor optically complex waters. Water column effects and Support Vector Machine (SVM) misclassifications were identified as limitations, indicating the need for additional improvement. The generated maps offer useful baseline information to direct management, restoration, and conservation initiatives for coral reefs.