River flow velocity is critical information for hydraulic and hydrological applications. Monitoring flow fields in near-plant waterways and river reach has wide engineering applications insustainable hydropower generation. For instance, exploring eco-hydraulic concerns such as fish migration, pollutant transport, and river erosion and understanding river floating debris are a few examples of practical applications. Due to complicated geometry and large volumes of natural flows, the measurement task using traditional methods (e.g., velocity propellers, acoustic Doppler velocimetry, with acoustic Doppler current profilers) usually requires extensive investigative work. The measurement procedure also requires contact with waterbody, thus avoiding its use in severe flood conditions. Image analysis approach allows the measurement task to capture the surface-water velocity distribution over a large outdoor area. The main objectives of this research are to (1) evaluate the feasibility of employing multiple cameras in a single measuring system to estimate the flow surfacevelocity and (2) improve the capability to use natural floating materials in river flow observations.

The properties of the camera system and particle tracking velocimetry (PTV) algorithm were investigated in a laboratory open channel flow measurement before being deployed for field measurements. The in situ camera calibration methods, which correspond to the two measurement situations, were used to mitigate the instability of the camera mechanism and camera geometry. The artificial tracer particles were deployed to seed the flows. Two photogrammetry-based PTV algorithms are presented regarding different types of employed seeding particles. The first algorithm uses the particle tracking method applied for individual particles, whereas the second algorithm employs correlation-based particle clustering tracking for clusters of small-size particles. The outcomes reveal that the method can offer a reliable and accurate assessment of 3D surface velocity.

In river surface velocity measurements, flow seeding is unavoidable in some situations where the water flow is clear, and there are no occurrences of floating materials on the surface. This part of the study focuses on the application of this technique for river velocity measurements using natural surfacefloating patterns. The use of a multiple-camera system provides the ability to perform 3D measurements on the river surface, including surface velocimetry and water surface reconstruction. The pattern-based tracking approach is used to adapt the performance of image measurements on different types of naturalfloating tracers. A comparison of pattern-based tracking with particle tracking reveals that these two approaches are consistent. An analysis of the characteristics of floating patterns is performed to understand their influences on standard deviation of measured velocity. Considerations on practicing image velocimetry in river flows are also discussed.