Improvement and Validation of Image Processing Methods for Floc Size, Settling Velocity, and Density.

The best techniques for characterizing floc properties combine in situ sampling with video-based systems with image analysis routines. The main advantage of using in situ video devices is the ability to measure floc size and settling velocity concurrently, with minimal disruption of hydrodynamics or change in floc size or settling. Concurrently measuring floc size and settling velocity allows for estimates of particle density thru theoretical or empirical Stokes based relationships. A major challenge in processing floc images is correctly identifying and sizing particles while omitting low contrast objects. This is especially challenging in estuarine and coastal environments, because natural flocs are not uniform in size or composition. This presentation introduces a new automated algorithm that efficiently and effectively identifies and sizes both large and small flocs, while rejecting out of focus objects. Field- and laboratory-based experiments were conducted to evaluate video-based size, settling velocity, and density estimates. The algorithm was implemented into the automated processing algorithm for the Particle Imaging Camera System (PICS, Smith and Friedrichs, 2015), which combines particle tracking and particle image velocimetry to measure settling velocity while accounting for background fluid velocity. Floc density is estimated by inverting a modified Stokes Law settling relationship. The accuracy of this approach was assessed with a (1) series of laboratory experiments using manufactured particles (both spherical and irregularly shaped) that cover a range of densities and sizes and (2) with natural mud aggregates of roughly known densities and size range. This paper provides validation for PICS video-based measurements of size and settling velocity, as well as density estimates inferred from size-settling relationships. Although the algorithm was created specifically for the PICS, it can be easily adapted for use with other video-based systems.