Detecting significant sediment motion in a laboratory flume using digital video image analysis

A video analysis method for monitoring sediment transport and sorting processes in a laboratory flume is presented. Video taken through the glass side-wall of a laboratory flume is captured using a digital CCD (charge-coupled device) camera and significant movements between individual frames are detected using image analysis. This method involves direct subtraction of the brightness numbers of pixels in sequential video frames, followed by thresholding to produce binarized images of significant change, above the inherent level of system noise. Experimental results showing dilation of a gravel framework and rapid infiltration of fines just prior to entrainment are discussed. Copyright (C) 2000 John Wiley & Sons, Ltd.