Digital Image Correlation Applied To The Calculation of The Out-Of-Plane Deformation Induced by The Formation of Roll Waves in A Non-Newtonian Fluid

Digital image correlation applied to
the calculation of the out-of-plane

deformation induced by the formation
of roll waves in a non-Newtonian
fluid
Alfredo Aranda
Nicolás Amigo
Christian Ihle
Aldo Tamburrino
Alfredo Aranda, Nicolás Amigo, Christian Ihle, Aldo Tamburrino, “Digital image correlation applied to the
calculation of the out-of-plane deformation induced by the formation of roll waves in a non-
Newtonian fluid,” Opt. Eng. 55(6), 064101 (2016), doi: 10.1117/1.OE.55.6.064101.
Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 4/2/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use

Optical Engineering 55(6), 064101 (June 2016)
Digital image correlation applied to the calculation of the

out-of-plane deformation induced by the formation of roll
waves in a non-Newtonian fluid
Alfredo Aranda,a,* Nicolás Amigo,a Christian Ihle,a,b and Aldo Tamburrinob,c
a
Universidad de Chile, Department of Mining Engineering, Tupper 2069, Santiago 8370451, Chile
b
Universidad de Chile, Advanced Mining Technology Center, Blanco Encalada 2002, Santiago 8370449, Chile
c
Universidad de Chile, Department of Civil Engineering, Blanco Encalada 2002, Santiago 8370449, Chile
Abstract. A method based on digital image correlation (DIC) is implemented for measuring the height of the roll
waves developed in a non-Newtonian fluid flowing on an inclined channel. A projector and a high-resolution
digital camera, placed vertically above the fluid surface, are used to project and record a random speckle pattern
located on the free liquid surface, where the pattern is deformed due to the developed roll waves. According to
the experimental geometry, the height of the roll waves associated to the out-of-plane deformation of the dots is
obtained through a quantitative relationship between the experimental parameters and the in-plane displace-
ment field in the flow direction. In terms of this, the out-of-plane deformation is found using a DIC criterion based
on the speckle comparison between a reference image without the deformed pattern and an image with a
deformed pattern. The maximum height of the roll waves computed with this technique is compared with
the height measured using a lateral camera, with both results differing by <10% over the set of experimental
instances. © 2016 Society of Photo-Optical Instrumentation Engineers (SPIE) [DOI: 10.1117/1.OE.55.6.064101]
Keywords: deformation of liquid surface; digital image correlation; height scalar field in roll waves; non-Newtonian fluids.
Paper 160342 received Mar. 9, 2016; accepted for publication May 16, 2016; published online Jun. 1, 2016.
1 Introduction discussed herein, some prominent techniques are the

The study and understanding of the surface waves developed Schlieren effect, employed for the visualization of the flow
in fluid flows are of essential importance for environmental of the natural convection in binary gas liquid-systems9 and
as well as industrial applications, such as tailing flows and the new development of an optical instrument which allows
mudslides. For example, flows of Newtonian fluids in the visualization of the discharge and distribution of fluids in
inclined channels frequently exhibit instabilities at their the ambient bottom water;10 the particle image velocimetry
free surface.1,2 In some cases, depending on the value of a or particle tracking used to measure the particle dynamics in
Reynolds number written out of the flow depth, the fluid vis- fluidized beds11 and the velocity field for an indoor airflow;12
cosity, and mean flow velocity, they evolve into roll waves. and the Fourier Transform Profilometry to measure the
Liu and Mei3 and later Huang and Garcia4 have modeled this deformation in and out-of-plane of a fluid.13,14 In this study,
type of flow in two-dimensions and the laminar flow regime a noninvasive technique to measure the height of roll waves
using boundary layer approximations when the (non- using digital image correlation (DIC) for out-of-plane meas-
Newtonian) fluid is a Bingham plastic, i.e., has a yield stress urement is implemented and tested in a laboratory flume dis-
and behaves as a Newtonian fluid in the limit of infinite shear charging a slurry. The method was first proposed by Pan
rate (see also for an extension to Herschel–Bulkley fluids5). et al.15 and was validated by means of static geometric fig-
In an infinitely long channel, two flow zones are distin- ures. Here, our aim is to apply this technique to identify the
guished using this constitutive model. One, near the bottom deformation of fluids in motion due to the propagation of roll
with a parabolic velocity distribution. A second zone, con- waves. In this regard, the DIC method has the crucial advan-
necting the parabolic velocity profile with the surface has a tage that it does not depend on the motion of the media, but
plug-shaped velocity profile,4 which is almost constant. on the exposure time of the camera. A prominent feature of
Roll waves in Bingham plastics depend on a critical this technique is that the height of the fluid can be measured
Reynolds and, additionally on the yield stress via the regardless its physical conditions, such as temperature or
Bingham number.6 Several studies have been carried out sediment transport. Moreover, the experimental setup is sim-
in the field of theses waves both in channels and unconfined ple, since only a high-resolution camera and the mechanical
geometries.7,8 The development of measurement techniques devices to fix it are required. As in particle image velocim-
for roll waves is, therefore, as relevant as the understanding etry, at least two images are needed in order to compare the
of their physical mechanism. Nowadays, there are many dif- deformed and undeformed projected pattern, which has a
ferent optical techniques to visualize or measure the proper- random distribution of dots. An important point to be out-
ties of fluid flows. For example, related to the approach lined is that the technique has a limit of precision of
0.01 pixel in the displacement field,16,17 thus allowing
*Address all correspondence to: Alfredo Aranda, E-mail: alfredo.arandan@

gmail.com 0091-3286/2016/$25.00 © 2016 SPIE
Optical Engineering 064101-1 June 2016 • Vol. 55(6)

Aranda et al.: Digital image correlation applied to the calculation of the out-of-plane deformation induced. . .
measuring with considerable spatial resolution. The tech- Δx1 þ x þ dp

nique, by definition, requires the projection of a mask on tan β ¼ : (4)
lp − h
EQ-TARGET;temp:intralink-;e004;326;752
a turbid surface, a requirement which is always fulfilled

when studying open channel (either confined or unconfined) Using Eqs. (1)–(2) and solving for h we obtain
mudflows. The article is organized as follows. In Sec. 2, the
principle of the DIC method and its equations are given. In Δx
Sec. 3, we show the experimental setup and the methodol- h¼
: (5)
dp 1
ogy. Then in Sec. 4, the results and discussion are presented.
dc
lc þ lp þ lp − l1c x þ l1p Δx
The conclusions are given in Sec. 5.
Following Pan et al.,15 we define x ¼ Mu, where M is the
image magnification and u is the local displacement field
2 Principle of Digital Image Correlation Method
in the x direction and at time t. Using the experimental set-
In the following section, we proceed to explain the basics of ting lp ¼ lc and M ¼ 1, Eq. (5) turns into the more conven-
the DIC method, as well as their main equations. ient form
u
2.1 Out-of-Plane Height of a Liquid Surface h¼d dp
: (6)
þ þ l1p u
c
lc lp
The DIC method measures the in-plane deformation and dis-
placement of the sample. However, it is possible to perform As it is seen in Eq. (6), the scalar height field is obtained by
out-of-plane measurements using geometric equations which measuring the experimental parameters and the displacement
depend on the displacement field.15 Figure 1 (adapted from field in the x direction, which is calculated from the DIC
Pan et al.15) shows the geometric scheme for a wave traveling algorithm.
in the −x direction, where the image parameters depend
solely on the spatial coordinates, valid for a given value
of time. In the scheme, Op and Oc represent the optic center
of the projector and the camera, respectively, which focus the 2.2 Digital Image Correlation Using a Speckle
x plane. The lp and lc dotted lines indicate the distance from Pattern
points Op and Oc to the x plane, respectively. Point o is DIC is an optical technique of a complete field, originally
chosen randomly at the border of the digital image. The dis- developed by Sutton et al.,18 with which in-plane displace-
tances dp and dc separate the projector and the camera to the ments and strains of the samples can be measured.15,19,20 This
origin, respectively, where dp þ dc represents the distance technique is based on projecting a pattern on the surface of
between the projector and the camera. In order to determine an object and capturing its initial and final deformed state,
the out-of-plane deformation, it is necessary to calculate the which are the called reference image and deformed image,
out-of-plane distance h, connecting the intersection of the respectively.21 Typically, these images are obtained through
segments Ōc B and Ōp A and the oX axis, as a function of a monochromatic CCD camera and they are digitally divided
variables that may be obtained directly from the geometry. in a region of interest (ROI) containing a finite number of
From Fig. 1, we have the following equations pixels22 with the purpose of optimizing the computational
correlation. The displacement field is calculated using a cor-
Δx1 relation algorithm which depends on the chosen criterion.23
EQ-TARGET;temp:intralink-;e001;63;361 tan α ¼ ; (1)
h In this study, the zero-mean normalized cross-correlation,
denoted as CZNCC , is shown23 following the result achieved
Δx2 by Blaber et al.24 for using an open-source MATLAB® algo-
EQ-TARGET;temp:intralink-;e002;63;319 tan β ¼ : (2) rithm. Optimal correlation using this algorithm requires val-
h
ues as close as possible to the unity
On the other hand, we have P
dc þ dp − ðΔx1 þ x þ dp Þ dc − Δx1 − x ½fðx̃ref i ; ỹref j Þ − f m ½gðx̃curi ; ỹcurj Þ − gm
tan α ¼ ¼ ; (3) CZNCC
EQ-TARGET;temp:intralink-;e007;326;283 ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P ffi;
lc − h lc − h ½fðx̃ref i ; ỹref j Þ − f m 2 ½gðx̃curi ; ỹc urj Þ − gm 2
(7)
where f and g are the reference and deformed image gray-

scale intensity functions at a specified location ðx; yÞ, respec-
tively. The functions f m and gm correspond to the mean
grayscale values of the reference and current subset, which
are indicated by
P
fðx̃ref i ; ỹref j Þ
fm ¼ ;
EQ-TARGET;temp:intralink-;e008;326;156 (8)
nðSÞ
P
gðx̃curi ; ỹcurj Þ
Fig. 1 Geometry employed to determine the out-of-plane gm ¼
EQ-TARGET;temp:intralink-;e009;326;109 : (9)
displacement. nðSÞ

Here, x~ ref i and y~ ref j are the x and y coordinates of an initial The subscript c denotes the central point of the analyzed
reference subset point, and x~ curi and y~ curj are the x 0 and y 0 image and u and v are the displacement fields of the x
coordinates of a deformed subset of points, which are and y directions, respectively. The deformation is character-
described by ized by the displacements u and v and their derivatives,
which are constant for a given subset.24 According to
∂u ∂u Pan,25 there exists a relationship between the ZNCC and
x̃curi ¼ xref i þ u þ ðxref i − xref c Þ þ ðyref j − yref c Þ; (10)
∂x ∂y
ZNSSD (zero-mean normalized sum of squared difference)

criterion as in
∂v ∂v
y~ curj ¼ yref j þ v þ ðx − xref c Þ þ ðyref j − yref c Þ: (11)
∂x ref i ∂y
The shape functions are identified from Eqs. (10) and CZNSSD ¼ 2ð1 − CZNCC Þ;
EQ-TARGET;temp:intralink-;e013;326;664 (13)
(11), which relate the deformed geometric parameters with
the reference geometric ones. From this functions, the
searching vector parameter, p is defined [Eq. (12)], which
is employed to determine the displacement field.24 where the CZNSSD criterion is usually employed because the
optimization of the ZNSSD coefficient is relatively easier
∂u ∂u ∂v ∂v T than that of the ZNCC coefficient.25 This one is defined
p ¼ uv
~ : (12)
∂x ∂y ∂x ∂y as follows:
8 92
X>
< fðx̃ref i ; ỹref j Þ − f m gðx̃curi ; ỹcurj Þ − gm
>
=
CZNSSD ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P − qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
P : (14)
>
: ½gðx̃curi ; ỹcurj Þ − gm 2 > ;
½fðx̃ref i ; ỹref j Þ − f m 2
7 deg and kept at this throughout the experiments. The cor-

The DIC displacement resolution can be extended to responding rheology is of a 14.58 Pa yield stress and of a
0.05 pixels by interpolating the correlation map to fractional 0.089 Pa · s plastic viscosity, using the Bingham model.
locations of the pixels.26 According to the resolution used in Two constant discharge volume flows have been made
this study, this value corresponds to 4 μm. Furthermore, the (Q1 ¼ 79.3 1.7 cm3 ∕s and Q2 ¼ 171.7 1.9 cm3 ∕s),
biquintic B-splines developed by Blaber et al.24 for the whereas in both conditions a uniform laminar flow was
MATLAB® graphical user interface is employed in order obtained. The experiments were conducted at standard tem-
to increase the calculation accuracy. perature and pressure conditions. On the metallic structure
On the other hand, the mask, represented by the speckle (2) a D3200 DSLR Nikon camera is placed to instantane-
pattern generated through an algorithm developed in ously capture the roll waves from above the flow, where
MATLAB®, is projected on the fluid surface. The pattern frames have been exposed at 1/30 s. The speckle pattern
has a spacial density of 7.3% which represents the space has been projected on the flow using a BENQ model
occupied by the white dots in the ROI whose dimensions MS517F data projector and the dot diameter was ∼4 pixels.
are 42 × 42 pixels. This value is calculated with On the plane base (1), a 190.0 × 15.0 × 1.7 cm3 melamine
channel, fitted transparent polycarbonate walls have been
As N s
ζ¼
EQ-TARGET;temp:intralink-;e015;63;286 ; (15) installed. This allowed video recording of the lateral profile
Ac by an independent means. For this purpose, a Sony model
DSC-WX300 digital camera has been used. The obtained
where As is the area of a single dot (all dots have the same images were digitalized to 8 bits and converted to grayscale.
area), N s is the number of dots which are in the subset, and From the lateral images, an intensity threshold to discern the
Ac is the area of the subset. It is important to emphasize that location of the free surface has been identified. On the other
with the chosen speckle pattern, a correlation coefficient hand, (4) is a wood base, supporting the 19-cm diameter and
>0.97 was obtained for all the images of this study. 40-cm height acrylic tank (5), used for loading purposes. The
latter is filled with a non-Newtonian liquid up to a 22 cm
3 Experimental Setup and Methodology height. The fluid is discharged into the channel using a PVC
A scheme of the experimental setup used in this study is tube (8), where two valves control the flow. The first one (7)
shown in Fig. 2. It consists of a 208.4 × 100.0 cm2 plane controls the outflow of the tank and the inflow into the chan-
base of melamine (impermeable to water), with a roughness nel is regulated using (8). The flow is interrogated at point
small enough to allow a uniform flow of bentonite. This dark (2), where both the data projector and the overhead camera
gray clay is completely turbid and, therefore, allowed the are installed. Finally, the fluid is recovered in the vessel (9).
projection of the speckle pattern onto it. In the present set The measuring process is as follows. First, the video
of experiments, a fixed concentration of bentonite of 11% recorded is initiated (both the side and top view cameras).
by weight has been considered, which is shown in Fig. 3. Then the valve for the outflow of the tank is opened and the
Also, the tilting angle of the channel has been set to flow is controlled with the inflow valve. After 5 s,

Fig. 2 Experimental setup. (a) Inclined channel on a melamine plane. (b) Projector and camera on the
metallics structure. (c) Side view of the camera and the projector.
corresponding to the time required in order to reach a steady 4 Results and Discussion
flow, eight images of different random roll waves are cap- The DIC method is used to identify the out-of-plane defor-
tured for a total of 18 waves. The process is repeated three mation of roll waves developed using the Q1 and Q2 . As the
times for the flow rate Q1 , obtaining a total of 24 images of first step, it is necessary to capture the speckle pattern pro-
the water–bentonite slurry. Then the process is repeated jected on the motionless fluid. Here, it is important to empha-
again for the flow rate Q2 , giving a total of 48 images. size that the fluid does not flow at the set angle because the
The two flow rates used during the experiments allowed to yield stress is greater than the gravitational force. This
study two different velocities of wave propagation, thus test- speckle pattern is shown in Fig. 4, and it is called the refer-
ing the proposed measurement algorithm at these two difference image, where it is noted that, due to the turbidity of the
ent conditions. The calibration of the system was done using bentonite, the pattern penetrates about ∼0.1 mm under the
three opaque objects with dimensions of 10 × 10 × 0.5 cm3 , surface of the fluid, thus deeming distortions due to changes
10 × 10 × 0.2 cm3 , and 10 × 10 × 0.8 cm3 (width, length, in optical properties between air and water irrelevant. In this
and height). First, the speckle pattern was projected over section, we present the results of the DIC method and the
the free-object surface. Then one of the objects was placed calculated maximum heights of the roll waves for Q1 and
on the surface and the pattern was again projected over it. Q2 . Only one flow of each case is analyzed in detail.
Thus, the measured height was obtained with the technique
giving the following results: 0.494, 0.202, and 0.796 cm,
respectively. From here, the factor of the calibration was cal-
culated, obtaining a value close to 1 (0.99).
Fig. 3 Fluid with a concentration of 11% of bentonite by weight. Fig. 4 Speckle pattern projected on the motionless fluid.

4.1 Digital Image Correlation Applied to the Q 1 Flow

Rate
The DIC method is applied to the Q1 flow rate. The speckle
pattern projected on a roll wave is shown in Fig. 5, where the
wave front is distinguished by analyzing the deformed pat-
tern shown in Fig. 4. The displacement field in the wavefront
direction is calculated using the DIC implementation pro-
posed by Blaber et al.24 To this purpose, the reference
image is correlated with the deformed image in Fig. 5,
allowing us to obtain the out-of-plane deformation presented
in Fig. 6, where the height is represented by the color dis-
tribution. The maximum calculated height for this roll wave
is obtained by adding the (constant) height field of the refer-
ence image (0.38 cm) to the calculated with the DIC tech-
nique, thus obtaining hDIC ¼ 0.97 cm, which is indicated
in the profile wave front shown in Fig. 7. This value is com-
Fig. 5 Speckle pattern and the front wave identified in the Q 1 flow
pared with the height measured using the lateral camera, hlat ,
rate. by capturing the wave front profile, which is observed in
Fig. 8(a). In order to calculate hlat , it is necessary to identify
the position of the free surface, which is done by converting
the image to a grayscale one and selecting one pixel with a
prescribed threshold intensity. Figure 8(b) shows an example
of the video-based measurement considering a threshold
value of 0.404 of the grayscale (scale from 0 to 1) corre-
sponding to the maximum height of the wave in the digital
image. Thus, hlat is equal to 0.92 cm, obtaining a percentage
difference of 4.65%. The above process is repeated for differ-
ent roll waves identified in the flow and their respective hDIC
and hlat are summarized in Table 1. As can be seen, the cal-
culated percentage differences are <10%, indicating that the
heights obtained with the DIC method are in good agreement
with those calculated using the lateral camera.
4.2 Digital Image Correlation Applied to the Q 2 Flow

Rate
The previous analysis is repeated for the roll waves identified
in the Q2 flow rate. The same reference image of Fig. 4 has
been used in this section. The speckle pattern projected on a
Fig. 6 Front wave and the height distribution of a roll wave in the Q 1 roll wave is shown in Fig. 9, where the front wave is indi-
flow rate.
cated. Again, this image is correlated with the reference
image to calculate the displacement field and thus obtain
Fig. 7 Wave front profile and its maximum height for a roll wave in the Q 1 flow rate.

Fig. 8 Wave front profile of a roll wave in the Q 1 flow rate. (a) Captured image and (b) its gray-scale
conversion.
Table 1 Maximum calculated h DIC and hlat heights and percentage differences for Q 1 .
Case 1 Case 2 Case 3
Wave h lat (cm) hDIC (cm) Δ% h hlat (cm) h DIC (cm) Δ% h h lat (cm) hDIC (cm) Δ% h
1 0.75 0.78 3.22 0.86 0.90 4.87 0.97 0.92 4.65
2 0.79 0.77 2.48 0.74 0.70 5.17 0.81 0.77 4.65
3 0.83 0.79 4.16 0.84 0.91 7.57 0.85 0.82 4.42
4 0.84 0.81 2.71 0.83 0.81 1.47 0.72 0.71 0.98
5 0.81 0.78 3.58 0.86 0.83 4.00 0.60 0.58 3.61
6 0.67 0.67 0.60 0.80 0.78 2.82 0.75 0.73 2.73
7 0.85 0.83 2.41 0.97 0.98 1.63 0.75 0.72 4.01
8 0.79 0.80 1.25 0.84 0.84 0.12 0.68 0.63 6.95
the out-of-plane deformation, as shown in Fig. 10. In this Fig. 12, we get hlat ¼ 0.83 cm. Hence, the percentage differ-
case, the maximum calculated height for this roll wave is ence 5.36%. This analysis is repeated for different roll waves
hDIC ¼ 0.88 cm, which is indicated in the profile wave identified for this flow. The results are presented in Table 2,
front shown in Fig. 11. Measuring the maximum height where the calculated percentage differences are <10%, as
using the captured image by the lateral camera, as seen in was already seen in the Q1 case.
Fig. 9 Speckle pattern and the front wave identified in the Q 2 flow Fig. 10 Front wave and the height distribution of a roll wave in the Q 2
rate. flow rate.

Fig. 11 Wave front profile and its maximum height for a roll wave in the Q 2 flow rate.
Fig. 12 Wave front profile of a roll wave in the Q 2 flow rate. (a) Captured image and (b) its gray-scale
conversion.
Table 2 Maximum calculated h DIC and hlat heights and percentage differences for Q 2 .
Case 1 Case 2 Case 3
Wave h lat (cm) h DIC (cm) Δ% h h lat (cm) h DIC (cm) Δ% h h lat (cm) h DIC (cm) Δ% h
1 0.79 0.79 0.38 0.74 0.72 2.65 0.71 0.72 1.25
2 0.78 0.79 0.63 0.78 0.83 5.21 0.77 0.77 0.00
3 0.81 0.83 2.65 0.68 0.68 0.29 0.83 0.84 1.43
4 0.83 0.88 5.36 0.74 0.72 2.21 0.83 0.83 0.24
5 0.74 0.75 1.21 0.72 0.71 1.41 0.77 0.78 2.04
6 0.77 0.78 2.04 0.78 0.77 0.26 0.78 0.82 4.98
7 0.76 0.81 5.59 0.75 0.77 2.84 0.70 0.74 5.03
8 0.75 0.79 4.45 0.78 0.79 1.77 0.82 0.79 3.79
5 Conclusions using images captured by a lateral camera. Both results were

In this study, the DIC was applied to analyze the out-of-plane in good agreement, which makes us presume that this tech-
deformation in turbid fluids, where it is possible to project a nique is capable not only of calculating deformations in sol-
speckle mask on the surface. The maximum height of differ- ids, as was already discussed by Gao et al.27 and Pan et al.,15
ent roll waves developed on an inclined channel was calcu- but also for calculating deformations in nontransparent fluids
lated and the results were compared to the height obtained by which are often encountered in different types of industries.

Furthermore, it was seen that the method was successfully 11. T. Hagemeier et al., “Estimation of particle dynamics in 2-D fluidized
beds using particle tracking velocimetry,” Particuology 22, 39–51
applied for two different volumetric flow rates. Hence, the (2015).
only limitation of capturing the images of the flow and 12. S. Fu, P. H. Biwole, and C. Mathis, “Particle tracking velocimetry for
indoor airflow field: a review,” Build. Environ. 87, 34–44 (2015).
then correlating them is the exposure time available in the 13. P. Cobelli et al., “Global measurement of water waves by fourier trans-
camera. However, it has the advantage that any flow rate form profilometry,” Exp. Fluids 46(6), 1037–1047 (2009).
can be measured depending on the exposure time of the cam- 14. E. Zappa and G. Busca, “Static and dynamic features of Fourier trans-
form profilometry: a review,” Opt. Lasers Eng. 50(8), 1140–1151
era. Finally, we emphasize that the DIC method is a remark- (2012).
ably useful technique for measuring deformations in fluids, 15. B. Pan et al., “Improved speckle projection profilometry for out-of-
as well as solids, because it is noninvasive and its application plane shape measurement,” Appl. Opt. 47, 5527–5533 (2008).
16. H. W. Schreier, J. R. Braasch, and M. A. Sutton, “Systematic errors in
does not depend on the physical conditions. digital image correlation caused by intensity interpolation,” Opt. Eng.
39(11), 2915–2921 (2000).
Acknowledgments 17. B. Pan et al., “Performance of sub-pixel registration algorithms in dig-
ital image correlation,” Meas. Sci. Technol. 17(6), 1615–1621 (2006).
AA thanks CONICyT PhD fellowship No. 21140180, and 18. M. Sutton et al., “Application of an optimized digital correlation
NA thanks CONICyT PhD fellowship No. 21151448. method to planar deformation analysis,” Image Vision Comput.
4(3), 143–150 (1986).
19. K. Wang, M. Gao, and A. Tieu, “Volume-grating digital speckle pat-
References tern interferometry for measurement of dynamic out-of-plane displace-
ment fields,” Measurement 43(4), 479–482 (2010).
1. B. Benjamin, “Wave formation in laminar flow down an inclined 20. H. Hu et al., “Digital speckle based strain measurement system for
plane,” J. Fluid Mech. 2(6), 554–573 (1957). forming limit diagram prediction,” Opt. Lasers Eng. 55, 12–21 (2014).
2. C. Yih, “Stability of liquid flow down an inclined plane,” J. Phys. 21. Y. H. Huang et al., “Shape measurement by the use of digital image
Fluids 6(3), 321–334 (1963). correlation,” Opt. Eng. 44(8), 087011 (2005).
3. K. F. Liu and C. C. Mei, “Slow spreading of Bingham fluid on an 22. J. Zhao et al., “Initial guess by improved population-based intelligent
inclined plane,” J. Fluid Mech. 207, 505–529 (1989). algorithms for large inter-frame deformation measurement using dig-
4. X. Huang and M. H. Garcia, “A perturbation solution for Bingham- ital image correlation,” Opt. Lasers Eng. 50(3), 473–490 (2012).
plastic mudflows,” J. Hydraul. Eng. 123(11), 986–994 (1997). 23. B. Pan, “Recent progress in digital image correlation,” Exp. Mech.
5. X. Huang and M. H. Garcia, “A Herschel–Bulkley model for mud flow 51(7), 1223–1235 (2011).
down a slope,” J. Fluid Mech. 374, 305–333 (1998). 24. J. Blaber, B. Adair, and A. Antoniou, “Ncorr: open-source 2-D digital
6. N. J. Balmforth and J. J. Liu, “Roll waves in mud,” J. Fluid Mech. 519, image correlation MATLAB software,” Exp. Mech. 55(6), 1105–1122
33–54 (2004). (2015).
7. Z.-Y. Wang, “Free surface instability of non-Newtonian laminar 25. B. Pan, H. Xie, and Z. Wang, “Equivalence of digital image correlation
flows,” J. Hydraul. Res. 40(4), 449–460 (2002). criteria for pattern matching,” Appl. Opt. 49, 5501–5509 (2010).
8. E. Shapar, E. Kalaidin, and E. Demekhin, “On the solitary roll-waves 26. Z. Liu et al., “Transmission-speckle correlation for measuring dynamic
on an inclined plane and their stability,” Thermophys. Aeromech. deformation of liquid surface,” Opt. Lasers Eng. 65, 110–116 (2015).
13(1), 75–83 (2006). 27. J. Gao, W. Xua, and J. Geng, “3-D shape reconstruction of teeth by
9. A. Okhotsimskii and M. Hozawa, “Schlieren visualization of natural shadow speckle correlation method,” Opt. Lasers Eng. 44(5), 455–465
convection in binary gasliquid systems,” Chem. Eng. Sci. 53(14), (2006).
2547–2573 (1998).
10. V. Karpen, L. Thomsen, and E. Suess, “A new Schlieren technique
application for fluid flow visualization at cold seep sites,” Mar. Biographies for the authors are not available.
Geol. 204(12), 145–159 (2004).

Digital Image Correlation Applied To The Calculation of The Out-Of-Plane Deformation Induced by The Formation of Roll Waves in A Non-Newtonian Fluid

Uploaded by

Copyright:

Available Formats

Digital Image Correlation Applied To The Calculation of The Out-Of-Plane Deformation Induced by The Formation of Roll Waves in A Non-Newtonian Fluid

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Digital Image Correlation Applied To The Calculation of The Out-Of-Plane Deformation Induced by The Formation of Roll Waves in A Non-Newtonian Fluid

Uploaded by

Copyright:

Available Formats

Digital image correlation applied to

the calculation of the out-of-plane

Downloaded From: https://www.spiedigitallibrary.org/journals/Optical-Engineering on 4/2/2018 Terms of Use: https://www.spiedigitallibrary.org/terms-of-use