Title: Real-Time Tracking
1Real-Time Tracking
Axel Pinz Image Based Measurement Group EMT
Institute of Electrical Measurement and
Measurement Signal Processing TU Graz Graz
University of Technology http//www.emt.tugraz.at
/tracking http//www.emt.tugraz.at/pinz axel.pin
z_at_tugraz.at
2Defining the Terms
- Real-Time
- Task dependent, in-the-loop
- Navigation on-time
- Video rate 30Hz
- High-speed tracking several kHz
- Tracking
- DoF Degrees of Freedom
- 2D images, videos ? 2 / 3 DoF
- 3D scenes, object pose ? 6 DoF
3Example High-speed, 2D
4Applications
- Surveillance
- Augmented reality
- Surgical navigation
- Motion capture (MoCap)
- Autonomous navigation
- Telecommunication
- Many industrial applications
5Example Augmented Reality
ARToolkit, Billinghurst, Kato, Demo at ISAR2000,
Munich http//www.hitl.washington.edu/research/sh
ared_space/download/
6AgendaStructure of the SSIP Lecture
- Intro, terminology, applications
- 2D motion analysis
- Geometry
- 3D motion analysis
- Practical considerations
- Existing systems
- Summary, conclusions
72D Motion Analysis
- Change detection
- Can be anything (not necessarily motion)
- Optical flow computation
- What is moving in which direction ?
- Hard in real time
- Data reduction required !
- Interest operators
- Points, lines, regions, contours
- Modeling required
- Motion models, object models
- Probabilistic modeling, prediction
8Change Detection
Pinz, Bildverstehen, 1994
9Optical Flow (1)
Brox, Bruhn, Papenberg, Weickert ECCV04 best
paper award
- Estimating the displacement field
- Assumptions
- Gray value constancy
- Gradient constancy
- Smoothness
- ...
- Error Minimization
10Optical Flow (2)
Brox, Bruhn, Papenberg, Weickert ECCV04 best
paper award !! Not in real-time !!
11Interest Operators
- Reduce the amount of data
- Track only salient features
- Support region ROI (region of interest)
Feature in ROI
Edge / Line
Blob
Corner
Contour
122D Point TrackingUniv. Erlangen, VAMPIRE,
EU-IST-2001-34401
- Corner detection ? Initialization
- Calculate cornerness c
- Threshold ? sensitivity, of corners
- E.g. Harris / Plessey corners in ROI
- Cross-correlation in ROI
132D Point TrackingUniv. Erlangen, VAMPIRE,
EU-IST-2001-34401
14Edge TrackingRapid 95, Harris, RoRapid 95,
Armstrong, Zisserman
15Blob TrackingMean Shift 03, Comaniciu, Meer
16Contour TrackingCONDENSATION 98-02, Isard,
Toyama, Blake
17CONDENSATION (2)
- CONditional DENSity propagATION
- Requires a good initialization
- Works with active contours
- Maintains / adapts a contour model
- Can keep more than one hypothesis
18AgendaStructure of the SSIP Lecture
- Intro, terminology, applications
- 2D motion analysis
- Geometry
- 3D motion analysis
- Practical considerations
- Existing systems
- Summary, conclusions
19Geometry
- Having motion in images
- What does it mean?
- What can be measured?
- Projective camera
- Algebraic projective geometry
- Camera calibration
- Computer Vision
- Reconstruction from uncalibrated views
- There are excellent textbooks
- Faugeras 1994, HartleyZisserman 2001, Ma et
al. 2003
20Projective Camera (1)
- Pinhole camera model
- p (x,y)T is the image of P (X,Y,Z )T
- (x,y) ... image-, (X,Y,Z) ... scene-coordinates
- o ... center of projection
- (x,y,z) ... camera coordinate system
- (x,y,-f) ... image plane
x
x
P(X,Y,Z)
f
o
z
y
y
Z
X
p(x,y)
Y
21Projective Camera (2)
- Pinhole camera model
- If scene- camera-coordinate system
X
f
o
Z
x
22Projective Camera (3)
- Frontal pinhole camera model
- (x,y,f) ... image plane
- Normalized camera f1
P(X,Y,Z)
f
x
x
p
o
z
y
y
Z
X
Y
23Projective Camera (4)
- real camera
- 5 Intrinsic parameters (K)
- Lens distortion
- 6 Extrinsic parameters (M R, t)
- ? arbitrary scale
24Algebraic Projective Geometry SempleKneebone
52
- Homogeneous coordinates
- Duality points ? lines
- Homography H describes any transformation
- E.g. image ? image transform x Hx
- All transforms can be described by 3x3 matrices
- Combination of transformations Matrix product
Translation Rotation
25Camera Calibration (1)
- Recover the 11 camera parameters
- 5 Intrinsic parameters (K fsx, fsy, fs?, u0, v0)
- 6 Extrinsic parameters (M R, t)
- Calibration target
- At least 6 point correspondences ?
- System of linear equations
- Direct (initial) solution for K and M
26Camera Calibration (2)
- Iterative optimization
- K, M, lens distortion
- E.g. Levenberg-Marquart
- Practical solutions require more points
- Many algorithms Tsai 87, Zhang 98, Heikkilä 00
- Overdetermined systems
- Robustness against outliers
- E.g. RANSAC
- Refer to Hartley, Zisserman, 2001
27What can be measured ...
- with a calibrated camera
- Viewing directions
- Angles between viewing directions
- 3D reconstruction more than 1 view required
- with uncalibrated camera(s)
- Computer Vision research of the past decade
- Hierarchy of geometries
- Projective oriented projective affine
similarity Euclidean
28AgendaStructure of the SSIP Lecture
- Intro, terminology, applications
- 2D motion analysis
- Geometry
- 3D motion analysis
- Practical considerations
- Existing systems
- Summary, conclusions
293D Motion AnalysisLocation and Orientation
head coord. system
R
t
scene coord. system
6 DoF pose in real-time
?Extrinsic parameters in real-time
303D Motion Analysis
- Tracking technologies, terminology
- Camera pose (PnP)
- Stereo, E, F, epipolar geometry
- Model-based tracking
- Confluence of 2D and 3D
- Fusion
- Kalman Filter
31Tracking Technologies (1)
- Mechanical tracking
- Magnetic tracking
- Acoustic time of flight
- Optical ? vision-based
- Compass
- GPS,
External effort required ! No self-contained
system
Allen, Bishop, Welch. Tracking Beyond 15
minutes of thought. SIGGRAPH01
32Tracking Technologies (2)Examples
Allen, Bishop, Welch. Tracking Beyond 15
minutes of thought. SIGGRAPH01
33Research at EMTHybrid Tracking HT
- Combine 2 technologies
- Vision-based
- Good results for slow motion
- Motion blur, occlusion, wrong matches
- Inertial
- Good results for fast motion
- Drift, noise, long term stability
- Fusion of complementary sensors !
Mimicks human cognition !
34Vision-Based Tracking More Terminology
- Measure position and orientation in real-time
- Obtain trajectories of object(s)
- Moving observer, egomotion inside-out
- Stationary observer outside-in Tracking
- Combinations of the above
- Degrees of Freedom DoF
- 3 DoF (mobile robot)
- 6 DoF (head tracking in AR)
35Inside-out Tracking
- monocular
- exterior parameters
- 6 DoF from ? 4 points
- wearable, fully mobile
corners
blobs
natural landmarks
36Outside-in Tracking
stereo-rig IR-illumination
- no cables
- 1 marker/device
- 3 DoF
- 2 markers 5 DoF
- 3 markers 6 DoF
devices
37Camera Pose Estimation
- Pose estimation Estimate extrinsic parameters
from known / unknown scene ? find R, t - Linear algorithms Quan, Zan, 1999
- Iterative algorithms Lu et al., 2000
- Point-based methods
- No geometry, just 3D points
- Model-based methods
- Object-model, e.g. CAD
38PnP(1)Perspective n-Point Problem
- Calibrated camera K, C (KKT)-1
- n point correspondences scene ? image
- Known scene coordinates of pi, and known
distances dij pi pj - Each pair (pi,pj) defines an angle ?
- ? can be measured (2 lines of sight, calibrated
camera) - ? constraint for the distance c pi
pi
pj
dij
xi
xj
?
c
39PnP (2)
pi
pj
dij
xi
xj
?
c
ui
uj
40PnP (3)
- P3P, 3 points
- underdetermined, 4 solutions
- P4P, 4 points
- overdetermined, 6 equations, 4 unknowns
- 4 x P3P, then find a common solution
- General problem PnP, n points
41PnP (4)
- Once the xi have been solved
- project image points ? scene
- pi xi K-1 ui
- find a common R, t for pi ? pi
- (point-correspondences ? solve a simple system
of linear equations)
42Stereo Reconstruction
z
Elementary stereo geometry in canonical
configuration 2 h baseline b Pr - Pl
disparity d There is just column
disparity Depth computation
P(x,y,z)
Pz
y
Cl
Cr
x
f
Pl
Pr
h
h
x0
xr0
xl0
43Stereo (2)
- 2 cameras, general configuration Epipolar
geometry
X
Y
ur
ul
v
ll
lr
Cl
Cr
el
er
44Stereo (3)
- Uncalibrated cameras Fundamental matrix F
- Calibrated cameras Essential matrix E
- 3x3, Rank 2
- Many Algorithms
- Normalized 8-point Hartley 97
- 5-point (structure and motion) Nister 03
45Model-Based Tracking
- Confluence of 2D and 3D Deutscher, Davison, Reid
01
463D Motion Analysis
- Tracking technologies, terminology
- Camera pose (PnP)
- Stereo, E, F, epipolar geometry
- Model-based tracking
- Confluence of 2D and 3D
- Fusion
- Kalman Filter
- General considerations
- Kalman Filter
- ? EMT HT project
47General Considerations
- We have
- Several sensors (vision, inertial, ...)
- Algorithms to deliver pose streams for each
sensor (at discrete times rates may vary
depending on sensor, processor, load, ...) - Thus, we need
- Algorithms for sensor fusion (weighting the
confidence in a sensor, sensor accuracy, ...) - Pose estimation including a temporal model
Allen, Bishop, Welch. Tracking Beyond 15
minutes of thought. SIGGRAPH01
48Sensor Fusion
- Dealing with ignorance
- Imprecision, ambiguity, contradiction, ...
- Mathematical models
- Fuzzy sets
- Dempster-Shafer evidence theory
- Probability theory
- Probabilistic modeling in Computer Vision
- The topic of this decade !
- Examples CONDENSATION, mean shift
49Kalman Filter (1)
Welch, Bishop. An Introduction to the Kalman
Filter. SIGGRAPH01
http//www.cs.unc.edu/welch/kalman
50Kalman Filter (2)
- Estimate a process
- with a measurement
x??n ... State of the process z??m ...
Measurement p, v ... Process and measurement
noise (zero mean) A ... n x n Matrix relates the
previous with the current time step B ... n x l
Matrix relates optional control input u to x H
... n x m Matrix relates state x to measurement z
51Kalman Filter (3)
52Kalman Filter (4)
53Kalman Filter (5)
http//www.cs.unc.edu/welch/kalman
54Kalman Filter (6)
http//www.cs.unc.edu/welch/kalman
55EMT Hybrid Tracker HT Project
- Ingredients of hybrid tracking
- Camera(s)
- Inertial sensors
- Feature extraction
- Pose estimation
- Structure estimation
- Real-time
- Synchronisation
- Kalman filter
- Sensor Fusion
56Hybrid Tracking
- 6 DoF vision-based tracker
- 6 DoF inertial tracker
- Fusion by a Kalman filter
57Structure and Motion
- Tracking Structure from Motion
58Research Prototype
- Tracking subsystem
- Visualization subsystem
- Sensors HMD
59HT Application Example
60AgendaStructure of the SSIP Lecture
- Intro, terminology, applications
- 2D motion analysis
- Geometry
- 3D motion analysis
- Practical considerations
- Existing systems
- Summary, conclusions
61Practical Considerations
- There are critical configurations !
- Projective geometry vs. discrete pixels
- Rays do not intersect !
- Error minimization algorithms required
- Robustness (many points) vs. real-time
- Outlier detection can become difficult !
- Precision (iterative) vs. real-time (linear)
- Combination of diverse features
- Points, lines, curves
- Jitter, lag
- Debugging of a real-time system !
62Existing Systems (1)
- VR/AR
- Intersense, Polhemus, A.R.T.
- MoCap
- Vicon, A.R.T.
- Medical tracking
- MedTronic, A.R.T.
- Fiducial tracker (Intersense)
- Research systems
- KLT (Kanade, Lucas, Tomasi)
- ARToolkit (Billinghurst)
- XVision (Hager)
63Existing Systems (2)
64Open Issues
- Tracking of natural landmarks
- First success in online structure and motion
- Nister CVPR03, ICCV03, ECCV04
- (Re-)Initialisation in highly complex scenes
- Usability !
65Future Applications
- Can pose (position and orientation) be exploited
? - What is the user looking at?
- Architecture, city guide, museum, emergency,
- From bulky gear and HMD ? PDA
- Wireless communication
- Camera(s)
- Inertial sensors ( compass, GPS, )
- Automotive !
- Driver assistance
- Autonomous vehicles, mobile robot navigation,
- Medicine !
- Surgical navigation
- Online fusion (temporal genesis, sensory modes, )
66Summary, Conclusions
- Real-time pose (6 DoF)
- 2D and 3D motion analysis
- Geometry
- Probabilistic modeling
- High potential for future developments
67Acknowledgements
- EU-IST-2001-34401 Vampire - Visual Active Memory
Processes and Interactive Retrieval - FWF P15748 Smart Tracking
- FWF P14470 Mobile Collaborative AR
- Christian Doppler Laboratory for Automotive
Measurement Research
- Markus Brandner
- Harald Ganster
- Bettina Halder
- Jochen Lackner
- Peter Lang
- Ulrich Mühlmann
- Miguel Ribo
- Hannes Siegl
- Christoph Stock
- Georg Teichtmeister
- Jürgen Wolf
68Further Reading
R. Hartley, A. Zisserman. Multiple View Geometry
in Computer Vision. Cambridge University Press,
2nd ed., 2003. Y. Ma, S. Soatto, J. Kosecka, S.
Shankar Sastry. An Invitation to 3D
Vision. Springer, 2004. B.D. Allen, G. Bishop,
G. Welch. Tracking Beyond 15 Minutes of
Thought. SIGGRAPH 2001, Course 11. See
http//www.cs.unc.edu/welch G. Welch, G.
Bishop. An Introduction to the Kalman Filter.
SIGGRAPH 2001, Course 8. See http//www.cs.unc.edu
/welch