A test bench to improve registration using RGB-D
sensors
François Pomerleau and Stéphane Magnenat and Francis Colas
and Ming Liu and Roland Siegwart
April 8, 2011 - European Robotics Forum, Västerås, Sweeden
Introduction Datasets Library+Tracker Conclusion
Introduction
I
Current works on icp/registration focus on limited
datasets/specific conditions
I
The best solution depends on the application
I
Yet there is a lack of benchmark/comparative studies
Solving this requires:
I
I
I
Common datasets
Versatile implementation
2/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Introduction
I
Current works on icp/registration focus on limited
datasets/specific conditions
I
The best solution depends on the application
I
Yet there is a lack of benchmark/comparative studies
Solving this requires:
I
I
I
Common datasets
Versatile implementation
2/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Introduction
I
Current works on icp/registration focus on limited
datasets/specific conditions
I
The best solution depends on the application
I
Yet there is a lack of benchmark/comparative studies
Solving this requires:
I
I
I
Common datasets
Versatile implementation
2/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Introduction
I
Current works on icp/registration focus on limited
datasets/specific conditions
I
The best solution depends on the application
I
Yet there is a lack of benchmark/comparative studies
Solving this requires:
I
I
I
Common datasets
Versatile implementation
2/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Common datasets
Datasets with ground truth, using a RGB-D sensor
3/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Dataset Acquisition Setup
I
Vicon
100 Hz, < 1 mm precision
I
Kinect
hand-held, 30 Hz, 160×120 points
I
Laptop
Thinkpad W510, Intel Core i7 Q820
http://www.asl.ethz.ch/research/datasets
access to FMA/Vicon through the courtesy of Prof. D’Andrea
4/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Experiments
I
translations on the three
axes, for about 10 s per axis
I
slow motion, speed of indoor
ground robots (0.3 m/s)
I
rotations on the three axes,
for about 10 s per axis
I
medium motion, speed of
agile robots (0.5 m/s)
I
a free fly motion over the
scene, for about 15 s
I
fast motion, a challenging
speed (1.3 m/s)
5/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Example of Acquisition: Free Fly
Video of free fly: slow, medium speed and fast
6/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Versatile implementation
Open-source library providing a modular icp chain
Tracker
7/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Tracker = Sequence of ICP + Key Frame
A
8/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Tracker = Sequence of ICP + Key Frame
A
8/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Tracker = Sequence of ICP + Key Frame
A
B
Minimum overlap
authorized
8/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Tracker = Sequence of ICP + Key Frame
A
B
Minimum overlap
authorized
8/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Tracker = Sequence of ICP + Key Frame
A
B
C
8/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Tracker = Sequence of ICP + Key Frame
F
A
E
B
C
8/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
D
Introduction Datasets Library+Tracker Conclusion
Tracking Performances
200
Old desktop PC with an Intel Xeon L5335
Recent laptop with an Intel Core i7 Q 820
Embedded system with an Intel Atom CPU Z530
Minimum number of points
Kinect frame rate
Frequency (Hz)
150
100
50
0
0
1000
2000
3000
4000
Number of point used
9/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
5000
6000
Introduction Datasets Library+Tracker Conclusion
Modular ICP chain
Reading
Reference
Data filter
Data filter
Data filter
Data filter
Data filter
Match
Outlier filter
Outlier filter
Error min.
Trans. checker
Trans. checker
10/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Iterations
Introduction Datasets Library+Tracker Conclusion
Typical ICP chain for Office Environments
Reading
Reference
Closest point reduction
Random sampling
Surface normal extraction
Kd-tree
Max distance
Median
Point-to-plane
Nb of iterations
Error
11/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Iterations
Introduction Datasets Library+Tracker Conclusion
Conclusion
Ongoing work toward standardisation of icp comparison
I
datasets with ground truth
I
open-source library providing a modular icp chain
Tracker as tech demo
12/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Conclusion
Ongoing work toward standardisation of icp comparison
I
datasets with ground truth
I
open-source library providing a modular icp chain
Tracker as tech demo
12/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Conclusion
Ongoing work toward standardisation of icp comparison
I
datasets with ground truth
I
open-source library providing a modular icp chain
Tracker as tech demo
12/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Conclusion
Ongoing work toward standardisation of icp comparison
I
datasets with ground truth
I
open-source library providing a modular icp chain
Tracker as tech demo
12/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors
Introduction Datasets Library+Tracker Conclusion
Live Demo and Questions
Enjoy!
Your questions are welcome!
13/13 — Pomerleau et al. — A test bench to improve registration using RGB-D sensors