Computer Vision Technique for Tracking Bed Load Movement

Abstract

An advanced image analysis system, called Khoros, was used to investigate the bed load movement of sediment particles in a laboratory flume. Incipient flow conditions prevailed throughout the experiments. Painted glass balls of identical diameter and density were used to simulate the sediment particles. They were uniformly placed on top of a tightly packed flat porous bed. Experiments were performed with two distinct surface packing configurations. A video camera was used to monitor their motion within a specified area of view. The resulting video record was converted to digital images using a frame grabber. These digital images were downloaded to a workstation for analysis. The outcome of this analysis provided quantitative information about the frequency of the entrainment of the glass beads, their displacement distance, and the mode of their motion. Such information, when used in conjunction with laser Doppler velocimeter measurements of the fluid velocity, can elucidate the physical mechanisms that are responsible for the entrainment of sediment. During the analysis of the tests, it was observed that the displacement of the beads was sporadic and occurred typically by rolling. The glass beads moved predominately along the flow direction, while on some occasions they were displaced in the transverse direction. For the two packing density tests that were examined, the minimum traveling distance in the longitudinal direction was found to be equal to one bead diameter and the maximum was equal to 10 bead diameters. In the transverse direction, the maximum particle traveling distance was equal to four bead diameters. Finally, it is shown that the existing imaging workspace can be used to accurately identify the displacements of small particles, which are typically encountered near incipient flow conditions and are not easily detectable with the bare eye. The imaging method described here is dynamic in nature and may prove to be a valuable tool for studying two-phase flows, as well as for visualizing flow structures taking place near the boundary in turbulent flows.