Comparative Assessment of Image-Based Velocimetry Techniques for River Flow Monitoring: Accuracy, Efficiency, and Practical Implication

الملخص

Accurate river flow monitoring is critical for water resource management, yet traditional methods face limitations in cost, spatial coverage, and real-time applicability. This study evaluates two open-source image-based velocimetry techniques, Large Scale Particle Image Velocimetry (LSPIV) from openrivercam and Particle Tracking Velocimetry (PTV) from tractrac, using field observation data from the Wesenitz and Freiberger Mulde rivers in Germany. Orthorectified video footage was processed to estimate surface velocities, validated against Acoustic Doppler Current Profiler (ADCP) measurements. Results demonstrate that openrivercam (LSPIV) achieved higher accuracy (R² = 0.57—0.62, RMSE = 0.11—0.23 m/s) for large-scale flow patterns, while tractrac (PTV) offered finer spatial resolution but was sensitive to noise. Computational efficiency favored tractrac with processing times 50-70% faster than another velocimetry tool. A sensitivity analysis revealed that time resolution and transect placement significantly influenced accuracy, whereas spatial resolution had a nuanced impact. These findings provide actionable insights for selecting velocimetry tools based on monitoring objectives, advancing non-invasive river flow monitoring.