- Article
Satellite-derived bathymetry (SDB) in inland waters remains challenging due to the strong spatial variability of optically active constituents and limited optical penetration depth. This study investigates the impact of a proposed Adaptive Spatial Filtering (ASF) approach and nonlinear regression models on bathymetric retrieval from PlanetScope imagery in two optically complex bays of Lake Dąbie, Poland. Single-band (SBA), multi-band (MBA), and band-ratio (BR) approaches were evaluated using linear regression (LR), generalized regression neural networks (GRNN), multilayer perceptron (MLP), and radial basis function neural networks (RBFNN). The results demonstrate that the proposed ASF method improves bathymetric reconstruction accuracy and surface quality. Within the threshold depth range, the average RMSE decreased from 0.195 m to 0.156 m, while coefficients of determination reached up to R2 = 0.914. Qualitative analyses further demonstrated that ASF improved the visual and geomorphological interpretability of SDB products and the overall quality of seabed reconstruction, enabling a more faithful representation of subtle morphological features and microtopographic structures. The SBA Red configuration provided the greatest effective mapping depth, reaching 2.76 m in Lubczyna Bay, whereas the MBA approach produced the most stable performance across varying depth ranges. GRNN achieved the highest accuracy within the threshold-depth range, while all neural-network models outperformed linear regression, except for the band-ratio configuration, which showed limited suitability for bathymetric retrieval in the study areas. The results indicate that local bio-optical conditions exert a stronger influence on SDB performance than depth alone and that no single spectral configuration is universally optimal for inland-water environments. The proposed methodology demonstrates the potential of combining adaptive filtering and nonlinear regression to enhance the applicability of satellite-derived bathymetry in optically complex inland waters for hydrographic surveys and environmental monitoring.
19 November 2026






















![Overview of the different housing structures according to Drexler et al. [25] by model year (a) and an evaluation of the distribution of the electric drive unit’s weight in kilograms (b).](https://mdpi-res.com/energies/energies-19-05324/article_deploy/html/images/energies-19-05324-g001-550.jpg)




