Drone-Based Bathymetry Modeling for Mountainous Shallow Rivers in Taiwan Using Machine Learning

The river cross-section elevation data are an essential parameter for engineering. However, due to the difficulty of mountainous surveys, existing bathymetry investigation techniques cannot be easily applied in a narrow and shallow field. Therefore, this study aimed establish model suitable areas utilizing unmanned aerial vehicle (UAV) equipped with multispectral camera machine learning-based gene-expression programming (GEP) algorithm. obtained images were combined total 171 water depth measurements (0.01–1.53 m) modeling. results show that coefficient determination (R2) GEP is 0.801, mean absolute error (MAE) 0.154 m, root square (RMSE) 0.195 m. performance has increased by 16.3% MAE, compared conventional simple linear regression (REG) algorithm, also lower retrieval both (<0.4 deep waters (>0.8 m). considerable degree accuracy could rivers or near-shore under similar conditions study.