Preliminary Result of Drone UAV Derived Multispectral Bathymetry in Coral Reef Ecosystem: A Case Study of Pemuteran Beach

Abstract

UAV-derived multispectral bathymetry is an alternative to creating a shallow water bathymetry map without a massive field survey. Multispectral UAV technology can be used for detailed scale identification scopes because it has better spatial resolution and relatively affordable cost. The UAV used in this study record the coastal area using four multispectral sensors, blue, green, red, and near-infrared bands. The UAV images are processed into point cloud information under the use of the Structure from Motion (SfM)-based algorithm with a spatial resolution of 0.075 m. Then the point cloud information is used to predict the water depth using the random forest algorithm. This research was conducted at Pemuteran Beach, Bali, Indonesia. We compared the performance of only spectral, cloud point, and the combination of cloud point – spectral information to predict the water depth. As a result, the cloud point – spectral based shows significant accuracy improvement compared with the spectral only approach that reaches ~1.5, ~2.5 m, and ~0.3m for R2, RMSE, and MAPE, respectively. So, the use of the SfM UAV technique can improve the common spectral-based SDB method.