21 August 2003 3D object recognition in TOF data sets
Author Affiliations +
Abstract
In the last years 3D-Vision systems based on the Time-Of-Flight (TOF) principle have gained more importance than Stereo Vision (SV). TOF offers a direct depth-data acquisition, whereas SV involves a great amount of computational power for a comparable 3D data set. Due to the enormous progress in TOF-techniques, nowadays 3D cameras can be manufactured and be used for many practical applications. Hence there is a great demand for new accurate algorithms for 3D object recognition and classification. This paper presents a new strategy and algorithm designed for a fast and solid object classification. A challenging example - accurate classification of a (half-) sphere - demonstrates the performance of the developed algorithm. Finally, the transition from a general model of the system to specific applications such as Intelligent Airbag Control and Robot Assistance in Surgery are introduced. The paper concludes with the current research results in the above mentioned fields.
© (2003) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Holger Hess, Holger Hess, Martin Albrecht, Martin Albrecht, Markus Grothof, Markus Grothof, Stephan Hussmann, Stephan Hussmann, Nikolaos Oikonomidis, Nikolaos Oikonomidis, Rudolf Schwarte, Rudolf Schwarte, } "3D object recognition in TOF data sets", Proc. SPIE 5086, Laser Radar Technology and Applications VIII, (21 August 2003); doi: 10.1117/12.486803; https://doi.org/10.1117/12.486803
PROCEEDINGS
8 PAGES


SHARE
Back to Top