Validation of a coastal zone bathymetric monitoring method based on the fusion of multimodal geospatial data acquired using UAV and USV sensors during a comprehensive measurement campaign

Abstract

Creating high-precision bathymetric models continues to be a major challenge, especially in complex environments where numerous factors restrict data gathering. To address these issues, this study introduces and evaluates a weighted average data fusion method. The approach was implemented during a comprehensive measurement campaign, in which weights were assigned to each dataset according to its estimated accuracy. All sensors, including UAV photogrammetry (Uncrewed Aerial Vehicle), SBES (Single-Beam Echo Sounder), MBES (Multibeam Echo Sounder), ALS (Airborne Laser Scanning), MLS (Mobile Laser Scanning), and GNSS RTK (Global Navigation Satellite System with Real-Time Kinematic positioning), were deployed simultaneously under consistent hydrometeorological conditions. The accuracy of the developed DTM (Digital Terrain Model) compared to actual measurements was assessed using RMSE (Root Mean Square Error) and MAE (Mean Absolute Error), which were 0.058 m and 0.025 m, respectively. These values indicate minimal differences between the interpolated and measured depths. The coefficient of determination reached 1.000, indicating a very strong agreement of the weighted average data fusion model with a power exponent of 1 to the measurement data. The depth estimation error for 68% of the data (R68 measure) was 0.020 m, while the error for 95% of the data (R95 measure) was 0.111 m. These results demonstrate not only the effectiveness of the proposed fusion method but also highlight the importance of synchronized, multimodal data collection. This approach enables the creation of detailed, high-accuracy bathymetric models with vertical accuracy values suitable for hydrographic and environmental monitoring applications in shallow inland waters.