Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
Digital Image Correlation: Overview of
Principles and Software
Correlated Solutions, Inc.
2D Image Correlation Fundamentals
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Outline
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Introduction
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Purpose
Deformation
measurement
▫ Full-field
measurement
▫ Non-intrusive
▫ Planar specimen only
▫ No out of plane motion
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Purpose
▫ Large range of size scales
(10-9 to 102 m.)
▫ Large range of time scales
(static to 200MHz)
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Monocular (cyclopean)
vision cannot determine
the size of objects
▫ Consequence: a 200%
isotropic deformation of
an object produces the
same image as if the
object was moved to one-
half its original distance
from the visual sensor
▫ We must assume the
object is planar, parallel
to and at a constant
distance from the visual
sensor during the entire
experiment
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Image Correlation Technique
▫ How? Given a point and its signature in the
undeformed image, search/track in deformed
image for the point which has a signature which
maximizes a similarity function
Time t Time t”Time t’
trackingtracking
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Image Correlation Technique
▫ In practice, a single value is not a unique
signature of a point, hence neighboring pixels
are used
▫ Such a collection of pixel values is called a
subset or window
Time t Time t’
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Image Correlation Technique
▫ The uniqueness of
each signature is
only guaranteed if
the surface has a
non repetitive,
isotropic, high
contrast pattern
▫ Random textures
fulfill this constraint
(speckle pattern)
Repetitive
Anisotropic
High-contrast
Non-repetitive
Anisotropic
High-contrast
Non-repetitive
Isotropic
Low-contrast
Non-repetitive
Isotropic
High-contrast
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Matching by DIC
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example
▫ The camera acquires 9x9 pixel images
▫ The specimen is marked with a cross-like pattern
Pixels
Image, on screen
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example
▫ White pixels are gray level 100
▫ Black pixels are gray level 0
▫ An image is a matrix of natural integers
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
Image, on screenImage, in memory
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example
▫ The specimen moves such that its image moves 1 pixel up
and right
100 100 100 0 100 100
100 100 100 0 100 100
0 0 0 0 0 0
100 100 100 0 100 100
100 100 100 0 100 100
100 100 100 0 100 100
100 0 0
100 0 0
0 0 0
100 0 0
100 0 0
100 0 0
0 0 0
0 0 0
100 0 0
0 0 0
0 0 0
0 100 100
0 0 0
0 0 0
100 100 100
Image after motion, on screenImage after motion, in memory
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example: we define a 5x5 subset in the reference image
(before motion)
▫ Problem: find where the subset moved (matching)
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
Image before motion
100 100 100 0 100 100
100 100 100 0 100 100
0 0 0 0 0 0
100 100 100 0 100 100
100 100 100 0 100 100
100 100 100 0 100 100
100 0 0
100 0 0
0 0 0
100 0 0
100 0 0
100 0 0
0 0 0
0 0 0
100 0 0
0 0 0
0 0 0
0 100 100
0 0 0
0 0 0
100 100 100
Image after motion
?
Subset
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Solution: check possible matches at several locations and use a similarity score (correlation function) to grade them
▫ Classic correlation function: sum of squared differences (SSD) of the pixel values (smaller values = better similarity)
22/
2/,
* )),(),((),,,( jvyiuxIjyixIvuyxCn
nji
Correlation function
Pixel coord., reference image
Displacement (disparity)
n: subset size (9 in our example)
Image before motion Image after motion
Pixel value at (x+i; y+j) Pixel value at (x+u+i; y+v+j)
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example: subset at (x;y)=(5;5), displacement candidate (u;v)=(-2;-2)
22
2,
* ))25,25()6,5(()2,2,5,5( jiIjiICji
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
Ima
ge
be
fore
mo
tion
100 100 100 0 100 100
100 100 100 0 100 100
0 0 0 0 0 0
100 100 100 0 100 100
100 100 100 0 100 100
100 100 100 0 100 100
100 0 0
100 0 0
0 0 0
100 0 0
100 0 0
100 0 0
0 0 0
0 0 0
100 0 0
0 0 0
0 0 0
0 100 100
0 0 0
0 0 0
100 100 100
Ima
ge
afte
r mo
tion
(x;y)=(5;5)
(u;v)=(-2;-2)
000,18)0100()1000()1000()1000()100100(
)00()1000()1000()1000()1000(
)00()1000()1000()1000()1000(
)00()1000()1000()1000()1000(
)0100()00()00()00()0100(
22222
22222
22222
22222
22222
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example: subset at (x;y)=(5;5), displacement candidate (u;v)=(1;1)
▫ Better correlation score than candidate (u;v)=(-2;-2) [18,000]Indeed it is the smallest score achieveable (perfect match)
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
100 100 100 100 100 100
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
0 0 0
Ima
ge
be
fore
mo
tion
100 100 100 0 100 100
100 100 100 0 100 100
0 0 0 0 0 0
100 100 100 0 100 100
100 100 100 0 100 100
100 100 100 0 100 100
100 0 0
100 0 0
0 0 0
100 0 0
100 0 0
100 0 0
0 0 0
0 0 0
100 0 0
0 0 0
0 0 0
0 100 100
0 0 0
0 0 0
100 100 100
Ima
ge
afte
r mo
tion
(x;y)=(5;5)
(u;v)=(1;1)
0)1,1,5,5( C
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ In reality, images are corrupted by some noise
▫ The SSD function will likely never be 0 for a perfect match
103 101 99 105 100 96
101 104 98 101 98 100
103 96 99 102 103 98
98 101 102 96 97 102
97 98 103 103 98 100
102 99 101 104 102 101
2 0 1
1 4 3
0 2 2
0 1 0
0 2 0
2 0 0
1 1 2
0 2 1
0 3 0
3 0 1
0 3 0
2 0 0
2 3 0
1 3 3
0 0 2
99 100 101 2 100 102
101 97 98 0 96 102
0 1 3 1 2 0
97 99 100 0 97 101
101 103 98 1 99 96
102 99 96 3 102 100
102 3 0
101 1 2
3 2 0
101 3 2
101 0 1
103 2 3
0 2 1
0 1 1
103 0 2
1 0 3
2 2 0
1 102 101
1 0 3
1 3 2
101 100 100
Image before motion Image after motion
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Principle
▫ Example
▫ The specimen moves such that its image moves 0.5
pixel to the right
▫ Need to interpolate the image at non-integer locations
103 101 99 55 100 96
101 104 98 51 98 100
103 96 99 52 103 98
98 101 102 46 97 102
97 98 103 53 98 100
102 99 101 54 102 101
52 0 1
51 4 3
49 2 2
52 1 0
51 2 0
48 0 0
1 1 2
0 2 1
0 3 0
3 0 1
0 3 0
2 0 0
2 3 0
1 3 3
0 0 2
Perfect match
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Photometric Mapping
▫ During image acquisition:
▫ Lighting conditions may change
▫ Sensor integration time adjusted
▫ Pattern may become lighter/darker when
expanded/compressed
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Photometric Mapping
▫ The photometric mapping is not
guaranteed to be an identity, hence the
DIC algorithm may have false matches
▫ Solution: model the photometric
transformation and use it to design a
robust correlation function
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Displacement Mapping
▫ General circumstances:
▫ The subset in the deformed image has changed
shape, e.g. a square initial subset is likely to be non-
square
Time t Time t’Specimen deformation
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Displacement Mapping
▫ Solution: model this displacement
transformation (called subset shape function)
and use it to define the deformed subset
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Image Interpolation
▫ Optimization algorithms require the criterion
to be continuous over its parameter space
▫ Images are discrete, hence we need to
reconstruct the continuous information by
means of interpolation
▫ B-splines are a good choice for that purpose
Raw image Bi-linear B-spline interp. Bi-cubic B-spline interp.
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Conclusions
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Summary
Acknowledgments
Conclusions
Purpose
Basic Idea
Principle
Measuring Displacem.
Introduction
Principle
Photometric Mapping
Displacement Mapping
Advanced Topics
Matching by DIC
Method
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Summary
▫ Matching process
▫ Photometric mapping
▫ Shape function
▫ Implications
Time t Time t”Time t’
trackingtracking
Three-Dimensional Digital Image Correlation
Overview
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Outline
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Principle
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Introduction
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Principle
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Purpose
▫ 3-D shape
▫ Strain map
▫ Full-field measurement
▫ Non-intrusive
▫ Any specimen shape
▫ Any motion
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Purpose
▫ Large range of size
scales
(10-9 to 102 m.)
3-D shape of letter „R‟ from “E PLURIBUS UNUM” on U.S. Penny coin using a Scanning Electron
Microscope
1X 50X 200X
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Basic Idea
▫ Much like human
vision, two imaging
sensors provide
enough information to
perceive the
environment in three-
dimensions
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Basic Idea
▫ DIC for stereo-
matching: two
3-D shapes
▫ DIC for
tracking:
relates the two
3-D shapes
through time
Temporal matching
(tracking)
Matching
by stereo-correlation
3-D
Matching by
stereo-correlation
3-DTime t
Time t’
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Principle
▫ Monocular (cyclopean)
vision cannot resolve
for the scale of
objects
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Principle
▫ Recovering the three-
dimensional structure
of the environment
using two imaging
sensors is called
stereo-triangulationObject coordinate
system
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Principle
▫ Stereo-
triangulation
requires computing
the intersection of
two optical rays
▫ Feasible only if
these rays are
formulated in a
common coordinate
system
▫ We need to model
and calibrate the
stereo-rig
? ?
?
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Calibration
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Principle
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Model
Perspective projection
Rigid transform Perspective projection
▫ Model: one extrinsic rigid transformation,
two intrinsic perspective projections
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Data Acquisition
▫ Acquire pairs of
images of a special
target undergoing
arbitrary motions
▫ Calibration is shape
measurement process
(bundle-adjustment)
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Parameter Estimation
▫ Dedicated pattern
recognition algorithm
Stereo RigTarget
3-D points
(approx.
known)
Pairs of
2-D points
(auto-recognized)
Transfer function
▫ Identification of the
transfer function Calibration target examples
Left: raw image, right: automatic pattern
recognition
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Parameter Estimation Techniques
▫ Bundle Adjustment
▫ No knowledge of calibration target required
aside from a scale.
▫ Shape of the target is measured during calibration.
▫ Typically, a relatively flat target is used for
initialization through a linear method.
▫ Generally regarded as the optimal method to
calibrate cameras.
▫ Complex to implement due to large equation
systems.
▫ Require block matrix solvers to efficiently invert
large equation systems.
▫ Can provide confidence margins on all
parameters estimated.
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Parameter Estimation Techniques
▫ There are only two constraints required
to accomplish calibration.
1. The calibration target must not deform
between images.
2. A distance between two points on the
target must be known.
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Principle
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
3-D Reconstruction
▫ 3-D Shape Measurement given a pair of images
Temporal matching
(tracking)
Matching
by stereo-correlation
3-D
Matching by
stereo-correlation
3-DTime t
Time t’
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
3-D Reconstruction
Two steps:
1. Stereo-correlation
2. Stereo-triangulation
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Tracking: Similar to 2-D DIC
▫ Allow to
relate 3-D
shapes
through
time
▫ Algorithms
similar to
2-D DIC
Temporal matching
(tracking)
Matching
by stereo-correlation
3-D
Matching by
stereo-correlation
3-DTime t
Time t’
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Combining everything
▫ The technology
behind 3-D DIC
▫ Stereo-correlation
▫ Tracking by
correlation
▫ Stereo-
triangulation
(needs
calibration)
Temporal matching
(tracking)
Matching
by stereo-correlation
3-D
Matching by
stereo-correlation
3-DTime t
Time t’
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Combining everything
▫ Example
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Strain Computation
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Principle
Strain Computation
Limitations, Benefits
Acknowledgments
Conclusions
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Principle
▫ Strains are computed from the measured
displacements of object points
(strains displacements)
3-D shape at time t
3-D shape at time t’
3-D displacement
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Principal
▫ Strains are only defined in the tangential
plane of the surface.
▫ The DIC data can be converted to a
triangular mesh.
▫ Since triangles are planar by definition, it
is simple to compute the strain on each
triangle.
▫ Same equations as found
in FEM.
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Smoothing
▫ The strains computed on each triangle are
noisy, unless the triangles are fairly large.
▫ The strains are normally smoothed using
low-pass filters.
▫ Low-pass filtering decreases spatial
resolution.
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Conclusions
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Principle
Strain Computation
Summary
Acknowledgments
Conclusions
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Purpose
Basic Idea
Camera Convention
Principle
Introduction
Model
Estimation
Calibration
Stereo Matching
Tracking
Stereo Tracking
Lens Distortion
Matching by DIC
Limitations, Benefits
Acknowledgments
Conclusions
Principle
Strain Computation
Summary
▫ Matching process – identical to 2D
▫ Combining matching with triangulation
▫ Combining triangulation with tracking