This paper compares model-based and artificial neural network approaches to computer vision based road following. Representative techniques of each approach, dynamic vision and radial basis function networks, were implemented in C/C++ on the same outdoor mobile robot using a commercially available workstation. Their performance was evaluated using the rules of the 1995 3rd Annual International Unmanned Ground Robotics Competition. Both methods were able to navigate a simple course; however, the model-based method performed more consistently and was found to have several practical advantages.
Implementation and comparison of two computer vision based road following approaches
Implementierung und Vergleich von zwei computervisionbasierten Straßenfolgeverfahren
1995
8 Seiten, 2 Bilder, 12 Quellen
Conference paper
English
Implementation and comparison of two computer vision based road following approaches
Automotive engineering | 1995
|Implementation and Comparison of Two Computer Vision Based Road Following Approaches
British Library Conference Proceedings | 1995
|Accurate Road Following and Reconstruction by Computer Vision
Online Contents | 2002
|