The Use of Imagery in Laboratory Experiments - Fluvial Remote Sensing for Science and Management - page 296

200

150

Flow direction

100

50

0

4.6mm

0 5 10 15 20

y - Transverse direction (pixels)

x (mm)

Figure 13.3 Example of three particles trajectories determined
using particle tracking algorithms. The time interval between
two successive positions is 0.004 s. The field of view is
23

Figure 13.4 Streamwise velocity of a single particle through
time as it moves downstream.

. × .

5mm 2

5 23 (particle diameter is approximately 50 pixels).

8 × 10 − 3

such that the particle diameter was about 50 pixels and
the field of view approximately 20 × 20 particle diameters.
At these conditions, the position of a dyed particle could
be determined with an accuracy of 0.05 mm. Particle
tracking was performed using algorithms developed using
the IgorPro data processing software (Lajeunesse et al.,
2010). The algorithm consisted of two steps. In the first
step each image was thresholded and binarised to create
an image in which the dyed particle was black and the
other particles were white. A particle detection algorithm
was applied in order to detect and localise the center of
each black particle. Lighting quality and the choice of
threshold criteria are crucial for this first step. The second
step consisted of reconstructing particle trajectories by
tracking the black beads through the sequence of images
(Figure 13.3). The data were then processed to calculate
streamwise and transverse particle velocities and their
distributions (Figures 13.4 and 13.5).
The main source of error in the experiments performed
by Lajeunesse et al. (2010) was oscillation of the water
surface which caused an apparent movement of the par-
ticles. This error can be corrected by determining the
apparent velocity of a particle at rest for each experiment.
Based on measurements of apparent velocity, a thresh-
old value is defined below which particle velocities are
discarded. Lajeunesse et al. (2010) had threshold values
between 10-30 mm/s, depending on the flow rate. The
same particle-tracking algorithm can be used to compute

6

4

2

0

0 200 400 600 800 1000

V x (mm/s)

Figure 13.5 Distribution of particle streamwise velocities ( Vx )
measured using particle tracking algorithms. The PDFs decrease
monotonically to zero and follow an exponential function.

lengths and durations of particle step-lengths. To do
this, the size of the field of view had to be increased so
that it is sufficiently larger than the characteristic particle
step-lengths.

13.2.2.2 Tracking of beads entrained in a quasi
two-dimensional channel
In this study PTV was performed from the side of a
quasi-two dimensional tilting flume (2 m long and 0.2 m
high) with transparent walls. The width of the flume was
only slightly wider than the width of the particle diame-
ters (6.5 mm). The goal of the experiments was to study
the motion of coarse spherical glass beads entrained by a