Iterative refraction-correction method on mvs-sfm for shallow stream bathymetry

Abstract. Extracting accurate bathymetric information from clear, shallow waters in complex riverine environments can be challenging, but it is crucial for many applications, such as hydraulic modeling, ecological monitoring, and sediment transport analysis. Multi-view stereo photogrammetry (MVS-SfM) has emerged as a promising technique for acquiring high-resolution bathymetric data from aerial imagery. However, the accuracy of MVS-SfM can be affected by various factors, including water refraction, which can distort the depth measurements. In this study, iterative Dietrich’s refraction-correction method is tested for extracting bathymetry from MVS-SfM in a complex riverine environment. Moreover, we proposed a workflow for applying the method using raster data files, which can be more readily available than point clouds. It also compared the obtained results with previous applications and evaluated them using statistical indices and ratios. For this case study, the multicamera refraction method produces bathymetric datasets with accuracies of ~0.019% of the flying height and precisions of ~0.07% of the flying height. This study contributes to increased confidence in exploiting aerial imagery for bathymetric mapping in photogrammetric procedures under field conditions and can facilitate the management and conservation of clear water systems.