Abstract
This chapter describes a bin-picking approach based on the Random Sample Matching approach. Objects are localized in 3D point clouds using 4D feature tables. After successful localization, a collision avoidance mechanism is introduced reducing the problem’s complexity by interpreting the point clouds as depth maps and using 2D image analysis. The chapter is concluded by the presentation of experiments demonstrating the potential of the approach.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
If this is not the case, the objects could be easily scanned.
- 2.
E.g. laser line scanners acquire depth information on equi-spaced points along a laser line which is translated across the scene, which leads to a regular grid of points. The same is true for depth cameras, which acquire depth information on a regular pixel grid.
- 3.
The system runs as demonstrator at the Institut für Robotik und Prozessinformatik since the end of 2012. Furthermore, the system was transferred into a prototypical work cell at Volkswagen Salzgitter where it also showed promising results.
- 4.
Time of Flight.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this chapter
Cite this chapter
Buchholz, D. (2016). 3D Point Cloud Based Pose Estimation. In: Bin-Picking. Studies in Systems, Decision and Control, vol 44. Springer, Cham. https://doi.org/10.1007/978-3-319-26500-1_3
Download citation
DOI: https://doi.org/10.1007/978-3-319-26500-1_3
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-26498-1
Online ISBN: 978-3-319-26500-1
eBook Packages: EngineeringEngineering (R0)