Benchmarking low-cost RPA photogrammetry against Terrestrial Laser Scanning for riverbank erosion monitoring
Abstract
Bank erosion in fluvial channels is a natural geomorphic process that can be significantly intensified by anthropogenic disturbances. This study assesses the contribution of bank erosion to morphodynamic changes in an urban channel through the integration of Terrestrial Laser Scanning (TLS) and low-cost Remotely Piloted Aircraft System (RPA) monitoring techniques. Both channel margins were surveyed over an extended temporal interval, enabling the generation of high-resolution Digital Elevation Models (DEMs) and derived Difference of DEMs (DoDs). Data processing was conducted using two vegetation-filtering approaches: a commercial software workflow and a customised MATLAB algorithm. Both methods provided consistent results, although the commercial software demonstrated superior performance in DoD analysis, while the MATLAB algorithm yielded cross-section profiles more comparable to TLS-derived data. Between 2019 and 2021, volumetric changes for Erosion Zone 01 were 13.55 m³ (commercial RPA), 8.32 m³ (MATLAB RPA), and 19.22 m³ (TLS); for Erosion Zone 02, the respective values were 73.22 m³, 84.69 m³, and 4.68 m³. TLS data exhibited greater stability and reliability in elevation measurement, whereas RPA-derived DEMs were more sensitive to variations in surface cover and vegetation density. These findings highlight the complementary potential of TLS and RPA for detailed quantification of bank erosion processes in urban environments.