Hough-based recognition of complex 3-D road scenes

Gian Luca Foresti; Carlo S. Regazzoni

doi:10.1117/12.135188

Translator Disclaimer

14 February 1992 Hough-based recognition of complex 3-D road scenes

Gian Luca Foresti, Carlo S. Regazzoni

Author Affiliations +

Proceedings Volume 1613, Mobile Robots VI; (1992) https://doi.org/10.1117/12.135188
Event: Robotics '91, 1991, Boston, MA, United States

Abstract

In this paper, we address the problem of the object recognition in a complex 3-D scene by detecting the 2-D object projection on the image-plane for an autonomous vehicle driving; in particular, the problems of road detection and obstacle avoidance in natural road scenes are investigated. A new implementation of the Hough Transform (HT), called Labeled Hough Transform (LHT), to extract and group symbolic features is here presented; the novelty of this method, in respect to the traditional approach, consists in the capability of splitting a maximum in the parameter space into noncontiguous segments, while performing voting. Results are presented on a road image containing obstacles which show the efficiency, good quality, and time performances of the algorithm.

Citation Download Citation

Gian Luca Foresti and Carlo S. Regazzoni "Hough-based recognition of complex 3-D road scenes", Proc. SPIE 1613, Mobile Robots VI, (14 February 1992); https://doi.org/10.1117/12.135188

ACCESS THE FULL ARTICLE