Image processing methods for in situ estimation of cohesive sediment floc size, settling velocity, and density

S. Jarrell Smith
Carl T. Friedrichs, Virginia Institute of Marine Science

Abstract

Recent advances in development of in situ video settling columns have significantly contributed toward fine-sediment dynamics research through concurrent measurement of suspended sediment floc size distributions and settling velocities, which together also allow inference of floc density. Along with image resolution and sizing, two additional challenges in video analysis from these devices are the automated tracking of settling particles and accounting for fluid motions within the settling column. A combination of particle track- ing velocimetry (PTV) and particle image velocimetry (PIV) image analysis techniques is described, which permits general automation of image analysis collected from video settling columns. In the fixed image plane, large-particle velocities are determined by PTV and small-particle velocities are tracked by PIV and treated as surrogates for fluid velocities. The large-particle settling velocity (relative to the suspending fluid) is determined by the vector difference of the large and small-particle settling velocities. The combined PTV/PIV image analysis approach is demonstrated for video settling column data collected within a dredge plume in Boston Harbor. The automated PTV/PIV approach significantly reduces uncertainties in measured settling velocity and inferred floc density.