The Use of Imagery in Laboratory Experiments - Fluvial Remote Sensing for Science and Management - page 304

Mean R1 = 156
Mean G1 = 165
Mean B1 = 147

250

200

Mean R2 = 106
Mean G2 = 127
Mean B2 = 87

150

Blue

Mean R3 = 59
Mean G3 = 104
Mean B3 = 39

100

50

0

300

200

100

Green

250

200

150

0

100

50

0

Red

Figure 13.13 Pixels from vegetation samples plotted in colour space. Each colour/symbol represents the class that the sample
corresponded to in the visual classification. The plot shows that the classes that were discerned visually do indeed fall into distinct
classes based on their colour value (Tal, 2008; Tal and Paola, in prep .).

(e.g., Kim and Paola, 2007; Hoyal and Sheets, 2009).
Another common technique consists of projecting single
or multiple laser line(s) onto the bed and photographing
it with a camera mounted at an oblique angle; vertical dis-
placements of the line are calibrated to real-world changes
in elevation (Figure 13.15; Leaf et al., 1993; Hasbargen and
Paola, 2000; Lague et al., 2003). These point-by-point
or line-by-line techniques are discontinuous in space
and some can be very time-consuming, requiring long
pauses in the experiment to perform each scan. Digital
photogrammetry is one method that provides a way of
measuring topography over a continuous surface (Chan-
dler et al., 2001; Lane et al, 2001; Turowski et al., 2006).
However, all these methods require a separate technique
and setup to measure flow depths. For example, combi-
nations of laser measurements and dye-density (discussed
above) have been used to simultaneously measure both
flow depth and bed topography (e.g. Huang et al., 2010).
Here we describe an optical method known as moir e
for acquiring measurements of both bed topography and
flow depth in laboratory experiments (Sansoni et al.,
1999). The moire projection method is part of a gen-
eral family of techniques that use the projection of
structured light to measure relief (Patorski, 1993) and
enables image-based non-contact measurements over a
continuous surface at very high spatial and temporal

resolutions. The moire method has been successfully
applied in metrology studies (Chiang, 1979), industrial
inspection (Sansoni, 2000), human body mapping for
medical diagnosis (Halioua, 1989; Kozlowski, 1997) and
art inspection (Bremand, 2007).
A moire method is based on projecting a fringe pat-
tern (also known as a grating or grid) on the bed and
analysing the deformation of the pattern caused by the
topography with respect to a fringe pattern projected on a
reference plane (i.e., a flat bed; Figure 13.16). The height
of the object (i.e., topography) is encoded in the phase
difference between the two patterns which is retrieved
through a Fourier transform or phase shifting algorithms
(Figure 13.17). While the mathematics behind the method
is rather complex, the good news is the methodology is rel-
atively easy to implement and user-friendly commercial
software that perform the calculations automatically are
available. For details about the theory and mathematical
operations readers should refer to Sansoni et al. (1999),
Pouliquen and Forterre (2002), and Limare et al. (2011).
Here we present only the basics behind the method, con-
siderations for implementing it, and an example from a
study of braided channels.
The simplest way to implement a moir´emethodcon-
sists of capturing an image (or series of images) of a single