Urban overland runoff velocity measurement with consumer-grade surveillance cameras and surface structure image velocimetry

Abstract Physically-based models are important tools for evaluating the hydraulic behaviour of urban drainage systems and, more specifically, assessing flood risk. While it is well known that such models should be calibrated and validated with monitoring data, overland runoff information is seldom available for this purpose. This study investigates the potential of using surveillance camera footage to measure surface flow velocity thanks to an LSPIV-based method called Surface Structure Image Velocimetry (SSIV). Seven real-scale experiments conducted in a specialized flood training facility were used to test the SSIV method under varied and challenging conditions. SSIV performance was evaluated by benchmarking bulk (mean) velocity against that measured by a conventional sensor array. In the best conditions tested, SSIV and conventional flow sensors differed by only 1.7% (0.1% standard deviation). While the method proved sensitive to light conditions, our results suggest that infrared lighting could be used to increase measurement consistency. Our study concludes that for measuring overland flow velocity in urban areas, surveillance and traffic cameras can be considered as a low-maintenance and easy-to-install alternative to conventional sensor systems.