Insight Tutorial

Particle Image Velocimetry
INSIGHT 4G TM
Data Acquisition, Analysis,

and Display Software Platform
Tutorial Guide
P/N 6005115, Revision A

June 2011
Manual History
This is a manual history of the INSIGHT 4G tutorial guide

(Part Number 6005115).
Revision Date
A June 2011
ii
Contents
Manual History .......................................................................................... ii
CHAPTER 1 Data Acquisition............................................................... 1-1

Prerequisites ..................................................................................... 1-1
Choose your Experiment and Run.................................................... 1-2
Adjust Exposure ................................................................................ 1-3
Adjust Seeding Density ..................................................................... 1-4
Evaluating Timing ............................................................................. 1-5
Adjusting Timing Parameters ............................................................ 1-6
Acquire Data! .................................................................................... 1-6
CHAPTER 2 Calibration ........................................................................ 2-1

Prerequisites ..................................................................................... 2-1
Create an Experiment ....................................................................... 2-1
Create a Run ..................................................................................... 2-2
Adjust Camera and Lens .................................................................. 2-2
Camera Aiming ................................................................................. 2-3
Fine Tuning Focus ............................................................................ 2-3
Focused ............................................................................................ 2-5
Acquiring Calibration Image .............................................................. 2-6
Creating Calibration .......................................................................... 2-6
CHAPTER 3 Processing PIV Images ................................................... 3-1

Introduction ....................................................................................... 3-1
Starting Out ....................................................................................... 3-2
Looking at an Image ......................................................................... 3-3
Default PIV Processor .................................................................... 3-4
Example ............................................................................................ 3-8
INSIGHT 4G Demo Data .............................................................. 3-8
Default Processor with 64 x 64 pixel Interrogation Window Size
and 16 pixel Maximum Displacement (0.25 of 64) ........................ 3-9
Default Processor ........................................................................... 3-10
Point Processing ............................................................................. 3-11
Change SNR Value......................................................................... 3-12
Default Processor with Spot Offset ................................................. 3-14
Image Pre-Processor ...................................................................... 3-15
Processed Image ............................................................................ 3-16
Increase Spatial Resolution ............................................................ 3-17
32 x 32 Pixels .................................................................................. 3-18
Recursive Nyquist Processing ........................................................ 3-19
Recursive Nyquist with 64 Pixel Starting Spot and 32 Pixel
Final Spot..................................................................................... 3-20
Post-Processing .............................................................................. 3-21
Vector Field ..................................................................................... 3-23
iii
Mask ............................................................................................... 3-24
Processed Image ............................................................................ 3-26
Viewing Vector Field in TecPlot ...................................................... 3-27
Calibration ....................................................................................... 3-28
Final Process .................................................................................. 3-29
Final Plot ......................................................................................... 3-29
Conclusion ...................................................................................... 3-29
CHAPTER 4 AutoMapping Demo Notes.............................................. 4-1

Stereo Auto Mapping ........................................................................ 4-2
3D Vector Processing ..................................................................... 4-31
iv INSIGHT 4G Software Tutorial Guide

CHAPTER 1
Data Acquisition
Prerequisites
Hardware connected, configured, and powered on
Synchronizer
Laser
Camera
INSIGHT 4G software installed and loaded
Experiment set up
Flow
Seeding
Laser Light Sheet
Experiment and Calibration Created in INSIGHT 4G software
1-1
Choose your Experiment and Run
Right-click on the experiment containing your
calibration
1
Select Set Experiment as Current
Calibrations are stored in an Experiment
and shared between Runs
Choose a Run to store your data
To acquire data into an existing run, Right-
Click and choose Set Run as Current
OR 2
To create a new run click the New Run
button
1-2 INSIGHT 4G Software Tutorial Guide

Adjust Exposure
First, go to the Capture tab
Next, set Application to PIV, Exposure to
Synchronized, and Capture to Continuous. 2
Set Lasers to Low Power
Ensure that there are seed particles present in
your flow
Start with a low seeding density
3
Select the smallest aperture on

your lens, e.g., 32
Press Capture
2,3
Adjust laser power and lens
aperture to give bright, crisply-
focused particle images
2,3 If image is dim, increase
laser power or increase the
1 aperture in 1-click increments
until you get a bright image.
If you see saturated pixels,
reduce laser power or select
a smaller aperture (larger
f-number)
If you see large saturated
areas, immediately stop
capture
Try to get a bright particle
image with no saturation at
f/8
Data Acquisition 3-3

Adjust Seeding Density
Adjust the quantity of seed particles in
your flow
Aim for a dense field of particles, but avoid
overseeding
In overseeded flows, many particles
overlap
The example image shows good seeding
appropriate for processing with default
settings
You can use the Grid feature of

INSIGHT 4G software to evaluate
seeding density
Set the X and Y distance to match the

interrogation spot size you are using. The 1
default is 64x64
Each grid square should contain several

(5-15) particle images
If grid squares do not contain several

particle images, increased seeding is
necessary
The example image is well-seeded for a 2

64x64 spot size

Evaluating Timing
After adjusting exposure and seeding density, you need
to set up your timing parameters
Stop Capture
2
Set Application to PIV, Exposure to Synchronized, and
Capture to Single
Press Capture
The main timing parameter that affects your data is Delta T

To evaluate Delta T, look at particle displacements between Frame A and Frame B
INSIGHT 4G software has several tools to help
Click on a frame to
switch to it
Click anywhere in the

image and the position
is given below click on
the same particle in
frames A and B to
measure displacement
For the best quality data, aim for a maximum particle displacement in the x
or y direction of < the size of your interrogation region
For the default interrogation region size of 64x64 pixels, you want
pixel displacements of 16 or less
Additionally, particles should travel no more than of your laser light
sheet thickness in the z direction
If particle motion appears random, then your Delta T is much too long
If particles appear stationary, your Delta T is much too short
Data Acquisition 3-5

Adjusting Timing Parameters
To adjust Delta T:
Click on the Timing Setup button on
the capture tab
1
Enter a value for Delta T in the dialog
Click Apply
Click Close 2
After adjusting Delta T, capture another
single pair of images and evaluate the
new Delta T
Repeat until you have a satisfactory setting
for your flow
4
3
Acquire Data!
Now that you have optimized your focus, 2
illumination, and timing you are ready to
take data.
To capture a sequence of images:
Select sequence as your capture
type 1
Click on the sequence setup button
Set up your parameters 3
Click Capture
Click Save RAM Images to save your
4
images to disk

CHAPTER 2
Calibration
Prerequisites
Hardware connected, configured, and powered on
Synchronizer
Laser
Camera
INSIGHT 4G software installed and loaded
Experiment set up
Flow
Seeding
Laser Light Sheet
Create an Experiment
First, create an
experiment to group all of
your settings and data
1
2-1
Create a Run
Next, create a Run to
store the settings and
data for a specific run
1
2
Adjust Camera and Lens

Before taking data set up the camera and lens. Start
out by running the camera in free mode using 2
ambient light
Go to the capture tab
Set the Application to PIV, Exposure to Free, and
Capture to Continuous.
Click Capture
3

Camera Aiming
Remove any filters from
lens
Place an object (For
example, a TSI calibration
target) in your
measurement plane)
Adjust the lens to its largest
(smallest number) aperture,
usually f/2.8 for most
accurate focusing
If the camera saturates
(magenta pixels) then
reduce your aperture
Adjust camera aiming and
focus until the image in
INSIGHT 4G software is
sharp and covers your area
of interest
Fine Tuning Focus

Now you need to fine-tune your focus
Remove your focusing target from the
measurement plane
For this you will focus on seed particles
illuminated by your laser sheet
Calibration 2-3
First, Stop Capture
Next, Set Application to PIV, Exposure to
Synchronized, and Capture to Continuous.
Set Lasers to Low Power
3
1
Select the smallest aperture on

your lens, e.g., 32
Press Capture 1,3
5
If the image on the screen is
dim, increase the aperture in
1-click increments until you get
a bright image. Do not exceed
the aperture setting you found
4 in Camera Aiming
If you see saturated pixels, go
2 back one click
If the image at the aperture
setting from Camera Aiming
is still dim, increase laser
power
Adjust focus until particle
images are crisply focused

To achieve best focus, view the particle image
zoomed in to at least Full Size 1:1. 1
Position the scroll bars to view the center
of the image 3
Individual particles should be crisply defined
The example image to the right shows a
focused particle field at Full Size 1:1
magnification
Focused
Congratulations, your camera is now
focused on the measurement plane
defined by your laser light sheet. In
order to retain this accurate focusing:
Keep the camera in the same
location
Keep the camera focus at the
same setting
Changing lens aperture is okay
If you move the camera or adjust the
focus, you will need to focus on laser-
illuminated particles again
Calibration 2-5
Acquiring Calibration Image
Use an object with markings spaced a known distance
apart 2
TSI Calibration Target
Ruler
Use free continuous capture to position your
calibration target in the measurement plane.
Set the Application to PIV, Exposure to Free, and 3
Capture to Continuous. 4
Click Capture
Ensure that your calibration target is visible on screen
and located in your laser light sheet plane
1 5
Click Stop and Save RAM Images to end the capture
and save the captured image to disk.
Creating Calibration
Keep your calibration image open
Click on the processing tab
Select Create New under spatial
2
calibration

Click Measure
Click on the image at 2 points a known
distance apart
Enter the physical distance between the
points you chose in the Object Size 2
box
4
Click OK
1
You are ready to acquire data!
3
5
Calibration 2-7
(This page intentionally left blank)

CHAPTER 3
Processing PIV Images
Introduction
Interpreting PIV data can be a daunting task, given the number of
processing algorithms available. Even within one commercial software
package such as TSIs INSIGHT 4G software, there are over 100
possible combinations of the Grid Engine, Spot Mask Engine,
Correlation Engine, and Peak Engine, not to mention the choice of
interrogation window spot size, and the maximum displacement.
This tutorial is by no means meant to cover every possible
combination. Only experience, extensive study of the papers and
literature, and getting your hands dirty will give you the knowledge to
know which algorithms work best under which circumstances.
That being said, this tutorial is meant to give the PIV user a tool to
begin to understand a couple of the most commonly used algorithms
for processing PIV data, and an understanding of how they are
different, and when it is most beneficial to use them. Keep in mind that
the explanations presented here are based on the research and
studies found in published works, and therefore, please refer to the
references section at the end of the tutorial for more details on the
ideas presented herein.
3-1
Starting Out
When you first open
INSIGHT 4G software,
you will see the
experiment tree. This is
where all of the images
and vector fields are
organized.
To analyze an image,
select it from the RawData
folder within a Run folder,
and drag and drop it into
the active area (gray
region).
Click on the Processing
tab in the lower-left corner
to begin processing.

Looking at an Image
This section details a
couple of different ways
you can look at the
image.
Right-click anywhere
on the image and draw
a rectangle to zoom in
on that region. To go
back to full size, click
on the Size button
and select Fit to
Window
Click on the Tools
button and select LUT
Table. Here you can
move the slider bars
around and select
Linear or Log to
change the way the
image appears on the
screen.
Note: this does not change the raw data, only the way it appears on the screen. This is useful for
accentuating dim particles.
Click on Pseudocolor, then Change Colors to change the colorbar, or the saturation color. The
default saturation color is Pink, so that it is alarming and warns the user that the image is saturated,
but it can be changed to other colors as well. Here, change it to white.
Processing PIV Images 3-3

Default PIV Processor
Application:
PIV
Spatial Calibration:
Create New
Name it.
Click the Measured
tab, and enter the
size of an object in
the flow, then click
Measure and click on
2 points on the image
corresponding to the
size of the object.
Click OK.
Processing Mask
Create New
Name it.
Draw a rectangle on
the image, this is
where you will end up
getting vectors.
Click OK.
Processing
Create New
Name it.
Use all the defaults
Click OK.
Click Start

The default PIV processor in
INSIGHT 4G software uses
the settings seen to the right.
This processor is useful for
getting a general idea of what
is happening in your flow.
The next several sections
describe what each setting
does.
Nyquist Grid
Defines the spots to be processed. Spot offsets
are defined in the grid engine.
Much like CFD data, the area of interest must
first be divided up into a grid, of course, PIV data
is always arranged in a rectangular grid.
The Nyquist Grid plugin is used for the fastest
results; it is the classic PIV grid. It sets vectors
with the x-spacing equal to half Spot A width and
the y-spacing equal to half Spot A height. No
spot offsets are used and the processing uses
only a single pass.
This gives a vector grid with 50% spot overlap
fitting the Nyquist sampling criteria.

No Mask Example of Gaussian Mask:
This setting means that Toggling Between
you will not be using a Frame A and Frame B
mask.
The spot mask is for
modifying the image spots
before correlation.
An example of a mask is
the Gaussian Mask,
which multiplies each pixel
value in spot A by a
Gaussian weighing
function so that the spot is
bright in the center and
dark around the edges.
This weighting gives more
value to the center pixels
and less to the edge pixels
[This is useful since the
particles near the edge
have a higher probability of
leaving the interrogation
region, and thus we would
like to give them less
importance in determining
the correlation (particle
displacement)].
FFT Correlator
Computes the correlation
using Fast Fourier
Transform (FFT). The
spots must be square
powers of two and Spot A
must be the same size as
Spot B.
If the spot size is not
square, the Zero Pad
Mask is automatically used
to pad the non-square part
of the window for FFT
correlation.

Gaussian Peak
Locates the correlation map peak with sub-
pixel accuracy by fitting a 3-Point Gaussian
in X and 3-Point Gaussian in Y with the
Gaussian Peak Fit
peak pixel used in both.
Intensity
Pixels
Notice that the Starting Spot

Dimensions are 64 x 64 for
Spot A, and 64 x 64 for Spot B.
The correlation spots in A and
B do not necessarily have to
be the same size. Sometimes
Spot B is chosen to be larger
so that it is more likely to
match particles that have left
the Spot A interrogation region.
64 x 64 or 32 x 32 pixels is
probably the most common
interrogation region window
size. Of course this value
should be chosen so that there
are at least 9-10 particles in
the interrogation region. If you
find that there are fewer than
9 particles, you may need to
increase the interrogation
window size. (e.g., 128 x
128 pixels).
The Maximum Displacement refers to the maximum particle image displacement, in pixels, that
will be looked for in processing. The Peak Engines use this value to limit the correlation map
search area to max displacement pixels from the zero pixel. Some Correlation Engines, such as
Hart, use this value to limit the size of the correlation map to increase processing speed and
decrease memory usage.

Example
Now an example will be shown. You will use the default PIV processor to
get an idea of the flow field, and then optimize the settings for the fewest
number of spurious vectors.
INSIGHT 4G Demo Data

The demo image pair called
turbulenceJet000001 will be used
included with INSIGHT 4G
software under:
C:\Experiments3G\Experiment PIV Demo\TurbulenceJet\RawData

Default Processor with 64 x 64 pixel
Interrogation Window Size and 16 pixel
Maximum Displacement (0.25 of 64)
This is the result of
processing
turbulenceJet000001
with the Default PIV
processor.
Green Vectors are
good data that have
passed the signal to
noise ratio tests in the
processor.
Red Vectors are bad
vectors. They did not
pass the SNR or
Minimum Avg Intensity
tests.
Go to Tools button
and click on Vectors
>> Setup Here you
can choose whether or
not to show the Bad
vectors. You can also
change the vector
scaling and color.
If you had imported your image from a non-TSI camera, you could set the T value on this screen.
(For TSI cameras, T is saved in the *.tif header, and the value on this screen is ignored)

Default Processor
A few things to notice:
First, the dominant flow direction is from right to left.
The places where we are missing the most vectors, seems to be
along the right side, where the velocity would be at a maximum. This
tells us that perhaps
T was too large,
spot size is too small, or
maximum displacement is too small.
Since we cannot go back and recapture the images with a different T
value, we will do the best we can.
In practice, often times perfect PIV images are just not possible. If
they were easy, someone else would have already made the
measurements!
Lets try to figure out what the particle displacement is along the edges,
using Point Process.
Here the vector scale has

been increased from 5 to 20.
The data does not look too
bad, but some good vectors on
the far right are being lost (red
vectors that appear to be
following the flow, not random).
Now check if these are good
vectors.
Click on the Point button.
You are going to have the
software analyze one
interrogation spot at a time, to
get an idea of why the red
vectors are failing.
Click anywhere on the image.
For this case, click on the right,
where you are getting failing
vectors.

Point Processing
The Processing Monitor tells
us that this particular
interrogation spot failed due to
the PeakToNoisePeak, which
was only 1.47, this must be at
least 1.5 (the default value).
Try lowering the SNR
threshold, to see if the data
gets better.
Also notice, that for this vector,
the pixel displacement dX was
-5.4 and dY was -2.89 (Y is
inverted)
Click Point again to make
the Processing Monitor close.

Change SNR Value
Click on Setup next to
the processor.
Click on Plugin
Settings
The default settings of
the processor are here.
You can change these
values by clicking on a
line, and then changing
the value in the lower
left corner. Be sure to
press Apply after
making each change.
Here, lower the SNR
filter value to 1.30

By changing the SNR
Correlation Plot
value from 1.5 to 1.3, Correlation Peak
we are saying that the
correlation peak must 0.1
be at least 1.3 times 0.09

0.09-0.1
0.08
higher than the second 0.08-0.09
0.07
highest (noise) peak. 0.06
0.07-0.08
Noise Peak 0.06-0.07
0.05 0.05-0.06
0.04 0.04-0.05
0.03 29 0.03-0.04
25
0.02 0.02-0.03
21
0.01 17 0.01-0.02
0 13 0-0.01
31 29 9
27 25
23 21 5
19 17
15 13
11 9 1
7 5 3 1

Default Processor with Spot Offset
This helped slightly, but
there are still some bad
vectors.
Another way to recover
vectors is to introduce a
spot offset. Since you
know that the dominant
displacement is to the
left, you can offset the
Spot B to the left. As
you will recall from a
previous slide, the
offset was
approximately -5 in the
X direction. Go back to
the process setup,
Plugin Settings, and put
in an offset of -4 for
xSpotOffset.
When you process, the
vector fields gets
slightly better, but not
too much.
Now try an image pre-
processor.

Image Pre-Processor
If you zoom-in on the
image, each particle
seems to occupy only
1-2 pixels.
Ideally you want 2.5
pixels or more.
The best way to do this
is to recapture the
images, with a different
magnification, or
different seeding
particles. Since you
already have the
image, work with it
from here.
Add an image pre-
processor that will blur
the particles so that
they encompass a
larger area. This also
has the effect of
boosting the brighter
particles, and
diminishing the dimmer
particles.
Under Pre-
processing, select
Create New. Name this
one Blur
Select Image Filter,
then double-click it to
enter Gaussian, 5, and
0.6.

Shown here are the before and after snapshots.
Notice how the final image has particles which encompass
more than 1-2 pixels.
Processed Image
The vector field looks much better.
Now see if you can go to higher
spatial resolution

Increase Spatial Resolution
Instead of 64 x 64 pixels,
try 32 x 32 pixels.

32 x 32 Pixels
Here you will find that
you have more vectors
total, but also more bad
vectors. To determine
the number of bad
vectors, go to the
Info >> Vectors >>
Vector Statistics, to
find the Total Vectors
(3,544) vs. Bad Vectors
(177) or,
95.2% good, and
4.76% bad.

Recursive Nyquist Processing
Recursive Processing is a
multi-pass method, in
which the first pass is used
as an estimate for the spot
offset in the second
passes.
Go to the Processor
Setup
For Grid Engine, select
RecursiveNyquistGrid
Change the starting spot
size to 64 x 64 and the
final spot size to 32 x 32
pixels
Click Pass Validation
Settings to filter bad
vectors after the 1st pass.
Double-click Local
Validation and Vector
Conditioning
(Double click the icon once
it is in the Pipeline, to
change the default
parameters)
Click OK.
Go back to the Plugin
Settings and change
the xSpotOffset to -4 and
verify that the SNR filter is
still set to 1.3
Click OK.

Recursive Nyquist with 64 Pixel Starting
Spot and 32 Pixel Final Spot
The vector field looks
much better.
You now have 96.7% good
vectors.
In general, you want to
have at least 95 to 97% or
more good vectors before
running the Post
Processing (vector
validation) step.

Post-Processing
Under Post
processing, click
Create New and
give it a name.
Double click the Local
Validation and the
Vector Conditioning
icons to add them to
the pipeline.
Double-click the Local
Validation icon.
Here you have the
option to change the
validation method. Click
OK.
Double-click the Vector
Conditioning icon.
Here you have the
option to change the
Hole Filling method,
or have the option of
Smoothing (Low-pass
filtering) the data.
Leave the default
values.
Click OK.
Click OK.

Back on the Processing Tab, you can select
None for Processor, and Select only a Post-
Processor (since you already have the vector field,
you do not need to repeat this step, you only need
to do the post-processing step)
Click Start

Vector Field
Vectors that have been
interpolated are shown in
Yellow, original vectors are
shown in Green.
Change the vector colors.
Go to Tools >> Vectors >>
Setup, and change the
vector colors to White and
Pink, double-click the color
to change it.

Mask
Perhaps you do not want vectors covering the entire PIV
image. There are several options for processing only part of
the image.
Region of Interest Processing This is done by using the
left-mouse button to drag a region on the image. When you
start processing (with a processor and a post-processor), you
will end up with vectors only in this region.
Mask This option allows you to mask in (and mask out)
regions of the image that you want to process.

In the Processing Mask box, select
Create New Name it.
Here you can create a mask that is a
Rectangle, an Ellipse, a Polygon, or
any combination.
To have a Rectangle and an Ellipse
simultaneously, draw the Rectangle
first, click on the ellipse tab, then
hold the control button, and draw the
ellipse.
Create a Mask of the center part of
the jet then click OK.

Processed Image
Reprocess the image, with the Mask
Selected, the Processor selected,
and the Post-Processor selected.

Viewing Vector Field in TecPlot
Click on the TecPlot
icon button. The last
vector field that was
processed, is
automatically loaded
into the TSI-TecPlot
Macro, and displayed
on the screen.
Change the Contour
Type to Flood, and
the Arrow Color to
Black.
Click on the Arrow
tab and make the
vectors longer by
clicking on the
Longer button.
Note that the units are
in Pixels and
Pixels/Second. This is
because you have not
yet calibrated the PIV
image. It gives you an
idea of the pixel
displacement, but
typically physical units
are more useful. Lets
calibrate.
Tecplot is a registered trademark of Tecplot, Inc., Bellevue, WA, USA.

Calibration
Back in INSIGHT 4G software, clear the
vectors by clicking the Tools Button >>
Vectors >> Clear.
Click create new under the Spatial Calibration
box, and give the calibration a name.
The box at the right appears.
To calibrate the image, associate the pixel size
(after magnification through the camera lens)
with a physical size (typically in microns).
(Note: you cannot just enter the actual size of
the pixels here, unless the magnification is
exactly 1:1)
The most common way to calibrate the image is
to measure something in the field of view, for
example a ruler that has been placed in the
plane of the light sheet. Since you do not have
an image of a ruler for this example, assume
that the distance from the top of the image to
the bottom of the image is 120 mm.
Type 120 in the Object Size (mm) field.
Click the Measure button.
Click on the top of the image, then click on the bottom of the image. The distance between these 2
points in pixels will appear in the Object Size (pixels) field, and the um/pixels calibration factor will
appear above.
Note: It is often useful to Zoom-in on the image while clicking on the 2 points. This can be done
by clicking the Size button and selecting a magnified field of view. Additionally, to ensure a
straight line, it is often useful to draw the line near the edge of the image, using the edge as a
reference.
Click OK.

Final Process
You can now process
the image with the
calibration, mask,
processor, and post-
processor selected.
Click Start.
Final Plot
Click the TecPlot button (or if
TecPlot is already running, the plot
will automatically load).
Play with the settings, to Plot a flood
of Vorticity.
The version of TecPlot that comes
with INSIGHT 4G software is a full
version, so manipulations of the data
outside of the macro are okay.
Conclusion
Congratulations! You have processed a PIV image pair. Of course there is
a lot more to PIV data processing than was shown here, but hopefully this
has given you an idea of what steps should be followed, and what is
possible with PIV software.

(This page intentionally left blank)

CHAPTER 4
AutoMapping Demo
Notes
The AutoMap experiment was created as an example of how to use the

Stereo Auto Mapping features of INSIGHT software. This document gives
step by step instructions on how to optimize the calibration using the
Stereo Auto Mapping procedure.
The AutoMap experiment is included as part of the Experiment PIV Demo

data that is installed with INSIGHT software. It contains two runs
AutoMap and AutoMapWindow, and the CalAutoMap calibration images in
the settings folder. The target was not precisely aligned with the light sheet
in the calibration images. The AutoMap run has flow images that can be
used to measure the light sheet location with respect to the target for
optimizing the calibration. The AutoMapWindow run has flow images
captured with a 13 mm thick glass window inserted between the cameras
and the light sheet after the calibration. This is a similar situation to
calibrating on a bench top and then making measurements in a wind tunnel
through a window.
4-1
Stereo Auto Mapping
Perspective Calibration Scan Images

Open INSIGHT software and
set Experiment PIV
Demo\AutoMap as the current
run. Launch Perspective
Calibration from INSIGHT
main menu | Tools |
Perspective Calibration. Open
the calibration images with
Perspective Calibration main
menu.
In the Calibration Image

Analysis dialog select Images
button. Select the two
calibration Images.

Click the Define Target
button and select 2 Planes,
no Traverse and Same Side
of Light Sheet (+z). In the
Two Plane Grid dialog, use
the default settings. Click OK.
Click the Options button and set the Edge Pixels to 100.
Click OK.
AutoMapping Demo Notes 4-3

Set Threshold to 10, Batch Analyze.
The Create Calibration Dialog

opens when the images have
been scanned. Set X Order
and Y Order to 3, Click
Create.

Check to make sure that the calibration points in the left, right and fluid
images look correct. Close the Perspective Calibration dialog with the X in
the upper right corner. The calibration .cpt file, PIVPerspCal and
dewarp_PIVPerspCal file have all been created.

Dewarp the Images
In INSIGHT software, select
the PIV Tab and Condition
Setup. In the Realization
Conditioner drop-down box
select Dewarping Image
Processor.
Press the Left Camera Frame A

button. In the Dewarping Setup dialog
press the Browse button.

Press the Look in: down arrow
button and verify that the
Experiment PIV Demo\Settings
is the current folder. Navigate
to the Experiment PIV Demo
\Settings folder if necessary.
Select EXPERIMENT PIV

DEMOPIVPerCal and click
Open. Note that the calibration
files are named for the
experiment EXPERIMENT PIV
DEMO and not for the
calibration image files
CalAutoMap.

Click OK to accept the
defaults.
Enter a name for the dewarping image conditioner you are

creating. Click OK.
Repeat the Experiments PIV

Demo.PIVPerCal selection for
Left Camera Frame B, Right
Camera Frame A and Right
Camera Frame B.

Select the
Close button.
The dewarping
image
conditioner
you created
will be
selected as
the Image
Conditioner.
In the
Experiment
Tree Tab,
open the
Flow\RawData
folder and
select all the
images. Open
them by right-
clicking and
choosing
Open.
In the PIV Tab set Size to Fit
to Window. Open the Tools |
PIV/PTV LUT Table Press
the Log button and click
Remember Settings then
Close.

Press the
Apply
Condition
button to
dewarp the
images.
Images after
dewarping,
notice the
particle area is
now a
trapezoid.

Ensemble Dewarp Test Process
From the PIV tab select Processor Setup... press the
New button.
Type in a name for the processor and click OK.
Change to Ensemble Average Correlation Processor, in the

Processor selection list.

Select Processor Settings.
In the Ensemble PIV

Processor Setup set the spot
sizes to 128, Maximum
Displacement to 63. Check
Stereo Dewarp.

Click on the PluginSettings
button, and Select the
NoMask:MinAvgIntensity by
double clicking.
Uncheck the SNR Filter Enable to turn off the minimum

spot intensity filter validation. Press OK.

Press OK in the Plugin Settings dialog and OK in the PIV
Processor Setup dialog.
Click the Close button in the PIV/PTV Processor Setup
dialog. The processing parameters are saved.
In the PIV Tab press the

Start Process button.

Validate the Vectors
Press the Validation Setup button, select Left Camera and
click OK.
Select the Macro Tab and press Start Recording

Enter a validation Macro Name and click OK.
Select the Median tab and click

Validate.

Select the Std. Dev. tab and click
Validate.
Return to the Macro tab and press

Stop Recording.

Press Done.
In the PIV
Validation
select the
macro you just
recorded and
Apply
Validation.

To use this validation macro
during processing press the
Process Setup.. button,
Processor Settings and
select it as the Final
Validation macro. Close the
dialog with OK, and close in
the PIV\PTV Processor
Setup Dialog.

Stereo Auto Mapping
In the vector validation
dialog statistics you could
see that the U mean
displacement is -11 pixels.
This is the mean disparity
error caused by the
calibration target not being in
light sheet plane. To Auto
Map and correct the target
alignment error select the
INSIGHT software main
menu Tools | Perspective
Calibration In the
Perspective Calibration
Dialog select Create |
Create Calibration Press
the Select *.CPT Files
button.
Select the
CalAutoMapcpt files.

Press the Position Target
button.

Browse for the Dewarp
Calibration File and select
dewarp_Experiment PIV
Demo.PIVPerCal. This is the
calibration after dewarping.
Browse for the WarpKey File

and Select WarpKey.txt.

Press the Optimize button
and the Browse for the Warp
Vector File.
Navigate to the
Flow\Analysis folder and
select the last left vector file.
This is the ensemble
average vector field of the
four left realizations, and
click Open.

Press the Calculate Best Fit Plane
button to compute an estimate of the
light sheet position in the target
coordinate system. Click OK.
Press the Create to create

and overwrite the calibration
files with the adjusted light
sheet location in the target
coordinate system.

The updated
calibration is
displayed
graphically.
Close
Perspective
Calibration.
Create a Dewarp Test Macro

To check the results
of the light sheet
position adjustment
you could repeat the
raw image selection
and dewarping
process. To automate
with a batch file follow
these steps. From the
INSIGHT software
main menu select
Tools | Visual Macro
Programmer. Drag
and drop the
Dewarp_AutoMappi
ngDemo.PIVCond
and Ensemble_DewarpTest.PIVProc that were created in previous steps into the step list.

Press the Save button and enter and macro name, OK then
Close.
To play the macro,

open the raw images
in the experiment
tree.
In the PIV Tab select

the macro from the
macro file list.

Press the Execute
Selected Macro
Lighting Bolt icon.
The Macro dewarps

the raw images and
then performs the
Ensemble Dewarp
Test processing on
the dewarped
images.
This auto map
corrected vector field
has much smaller
mapping errors than
the original
calibration. The auto
mapping may now be
repeated with this
vector field to further
reduce the mapping
errors.

From the INSIGHT software
main menu select Tools |
Perspective Calibration
In the Perspective calibration
dialog select Calibrate Select
Calibration. The dialog open
with the cpt files and target
position from the first auto
mapping. Press the Position
Target button.
The Dewarp Calibration File and

WarpKey file are remembered and
do not have to be reselected. Press
Optimize.
The light sheet position from the

first optimization is displayed. The
Warp Vector File is the vector file
from the first auto mapping. Press
the Browse button.

Select the last left vector
field from the second
processing pass. You can
tell the processing pass by
the .P004. in the file name.
Press the Calculate Best Fit Plane

button. The light sheet position error
from this second pass vector field is
added to the first pass results. The
change to the light sheet position is
small. Press OK, and create to
update the calibration file. Close
Perspective Calibration.

Create a Dewarp Test Batch File
To increase the automation, create a batch file. In the Experiment
Tree select the raw images. Right-click and select Create Image
List. Enter a file name and press OK.
Right-click on an open area below the last experiment, and select

Create Batch File. Enter in a batch file name and press OK.
The batch file

appears below the
experiment list. Use
the INSIGHT main
menu |
Experiment/Run
Refresh Tree if
necessary. Open the
Settings folder in the
experiment tree. Click
on the
DisparityTestMacro.
mac file and drag on
drop it on the
Disparity_AutoMap.
batch file.
When the file is over the batch file, the

cursor changes. A dialog opens. The
dialog can be resized by dragging the
lower-right corner.
In the dialog select the image list from
the list of available image lists and
press OK.

The batch setup can also be
setup by dragging the image
list and selecting the
processing macro in the
dialog.
As many image list,
processing macro pairs as
you want may be loaded into
the batch processor, but you
only need one processing
step now.
Right-click on the batch file
and select Run to open the
image list, dewarp and
process the images. Press
the Run button to play the
batch file.
When the processing is
finished the image list closes
and the dialog message is set
to Batch Complete.
The vector field may be used to further optimize the calibration as was
shown in the macro section. The changes to the calibration will be quite
small.
Making Measurements through a Window

In some experiments it can be difficult or impossible to place the calibration
target in the flow area. The Stereo Auto Mapping procedure can be used to
compensate for the calibration shift due to imaging through the window.
The window will change the apparent target z axis location.

AutoMapWindow run has example images of the same flow viewed
through a 13 mm thick glass window. You can see how the calibration
changes by repeating the Stereo Auto Mapping process with the
AutoMapWindow images.
3D Vector Processing
Open the images in the experiment tree.
Set the Processing ROI

Select the ROI tab.
Select a
processing Region
Of Interest (ROI)
by placing the
cursor over one
corner of the ROI
area and pressing
down on the left
mouse button.
Hold the mouse
button down and
move the cursor to
the opposite
corner and
release.
Press the Save button and enter a ROI name and press OK.

Create a Stereo Processor
Press the PIV tab and Process Setup New.
Enter a processor name and press OK.
Set the Processor to PIV Processor.

Press the Processor
Settings button. Set the spot
sizes to 32, Maximum
Displacement to 12, and the
Correlation to PIV,
MedianStDev Final Validation.
Select the Experiment PIV
Demo 3D calibration file.
Deselect Keep Correlation
Maps for Validation if you are
not going to use RHP
validation to save memory.
Press OK and Close in the
PIV/PTV Processor Setup.
Create a Processing Macro for Stereo Processing

From the
INSIGHT
software main
menu select
Tools | Visual
Macro
Programmer
|. Drag and
drop the ROI,
and Processor
into the steps
list.

Save the Macro, press OK, then Close.
With the
images open
select the
Macro and
Macro Play
buttons.
Vector fields
are shown as
an image
overlay.

Batch Stereo Processing
The Stereo
PIV macro can
be used with
the
AutoMap_List
from the auto
mapping in a
batch file. In
the
Experiment
Tree tab right
click below the
list and accept
Create Batch
File.
Enter the batch name and press OK.

Open the
Settings and
drag the
image list over
the batch file
and drop.
There are now two macros in

the experiment, select the
StereoPIV_Macro and press
OK.

Right-click on
the batch file
and select
Run.
Click on Run to run the

batch file.

TSI Incorporated 500 Cardigan Road, Shoreview, MN 55126 U.S.A
USA Tel: +1 800 874 2811 E-mail: fluid@tsi.com Website: www.tsi.com
UK Tel: +44 149 4 459200 E-mail: tsiuk@tsi.com Website: www.tsiinc.co.uk
France Tel: +33 491 11 87 64 E-mail: tsifrance@tsi.com Website: www.tsiinc.fr
Germany Tel: +49 241 523030 E-mail: tsigmbh@tsi.com Website: www.tsiinc.de
India Tel: +91 80 41132470 E-mail: tsi-india@tsi.com
China Tel: +86 10 8251 6588 E-mail: tsibeijing@tsi.com
Singapore Tel: +65 6595 6388 E-mail: tsi-singapore@tsi.com
Contact your local TSI Distributor or visit our website www.tsi.com for more detailed specifications.
P/N 6005115 Rev. A Copyright 2011 by TSI Incorporated Printed in U.S.A

Insight Tutorial

Uploaded by

Document Information

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

Insight Tutorial

Uploaded by

Copyright:

Available Formats

Particle Image Velocimetry

Data Acquisition, Analysis,

P/N 6005115, Revision A

This is a manual history of the INSIGHT 4G tutorial guide

Manual History .......................................................................................... ii

CHAPTER 1 Data Acquisition............................................................... 1-1

CHAPTER 2 Calibration ........................................................................ 2-1

CHAPTER 3 Processing PIV Images ................................................... 3-1

CHAPTER 4 AutoMapping Demo Notes.............................................. 4-1

iv INSIGHT 4G Software Tutorial Guide

1-2 INSIGHT 4G Software Tutorial Guide

Select the smallest aperture on

Data Acquisition 3-3

You can use the Grid feature of

Set the X and Y distance to match the

Each grid square should contain several

If grid squares do not contain several

The example image is well-seeded for a 2

1-4 INSIGHT 4G Software Tutorial Guide

The main timing parameter that affects your data is Delta T

Click anywhere in the

Data Acquisition 3-5

1-6 INSIGHT 4G Software Tutorial Guide

Adjust Camera and Lens

2-2 INSIGHT 4G Software Tutorial Guide

Fine Tuning Focus

Select the smallest aperture on

2-4 INSIGHT 4G Software Tutorial Guide

2-6 INSIGHT 4G Software Tutorial Guide

2-8 INSIGHT 4G Software Tutorial Guide

3-2 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-3

3-4 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-5

3-6 INSIGHT 4G Software Tutorial Guide

Notice that the Starting Spot

Processing PIV Images 3-7

INSIGHT 4G Demo Data

C:\Experiments3G\Experiment PIV Demo\TurbulenceJet\RawData

3-8 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-9

Here the vector scale has

3-10 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-11

3-12 INSIGHT 4G Software Tutorial Guide

be at least 1.3 times 0.09

Processing PIV Images 3-13

3-14 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-15

3-16 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-17

3-18 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-19

3-20 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-21

3-22 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-23

3-24 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-25

3-26 INSIGHT 4G Software Tutorial Guide

Tecplot is a registered trademark of Tecplot, Inc., Bellevue, WA, USA.

Processing PIV Images 3-27

3-28 INSIGHT 4G Software Tutorial Guide

Processing PIV Images 3-29