Florian Pucher, Hubert Gattringer, Andreas Müller,
"Analysis of Feature Tracking Methods for Vision-Based Vibration Damping of Flexible Link Robots"
: Proceedings of the Austrian Robotics Workshop 2018, Seite(n) 25-30, 2018
Original Titel:
Analysis of Feature Tracking Methods for Vision-Based Vibration Damping of Flexible Link Robots
Sprache des Titels:
Englisch
Original Buchtitel:
Proceedings of the Austrian Robotics Workshop 2018
Original Kurzfassung:
Computer vision is often used in robotics where
image-based feature detection is an important part. The obtained features can be exploited, e.g. for path planning, process
monitoring or feedback control. In this paper the focus is on
vision-based vibration damping of robots with flexible links.
The measured values for control are obtained by extracting
image features. The required image processing framerate depends on the link dynamics. Image processing in general is a
computationally expensive task since the complexity for pixel
operations is of order O(n2). Efficient algorithms for online
feature tracking have to be used. In an experiment, image
processing is performed on a low cost computer and results
regarding the computational time are presented. The feature
detection performance is validated by results of the vision-based
vibration damping control.
Forschungs-