Satellite-derived bathymetry combined with Sentinel-2 and ICESat-2 datasets using machine learning

Most satellite-derived bathymetry (SDB) methods developed thus far from passive remote sensing data have required in situ water depth, limiting their utility areas with no data. Recently, new Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2) observations shown great potential providing high-precision a priori depth benefits range-resolved lidar. In this study, we propose combined active SDB method using only satellite An adaptive ellipse DBSCAN (AE-DBSCAN) algorithm is introduced to derive bathymetric ICESat-2 automatically adapt the terrain change complexity, then use these Sentinel-2 images help build model between reflectance (Rrs) depth. Three machine learning (ML) are employed, performances compared conventional empirical models presented. After that, results different Rrs band combinations effects without atmospheric correction on ML-based discussed. The showed that our AE-DBSCAN performs better than standard method, can achieve an overall RMSE of less 1.5 m St. Thomas traditional method. They also indicate obtain relatively correction, which helps improve processing efficiency by avoiding complex process.