Citation

Subirats P, Dupuis Y. Transp. Lett. 2015; 7(2): 114-117.

Copyright

(Copyright © 2015, Maney Publishing, Publisher Informa - Taylor and Francis Group)

DOI

10.1179/1942787514Y.0000000038

PMID

unavailable

Abstract

In this paper, the authors investigate the problem of detecting and counting motorcycles in real traffic flow with a single layer laser scanner. This work focuses on the use of simple signal processing techniques to achieve a good performance in real time. The reported results outperform state-of-the art results. First, the authors report a 99·2% correct classification rate on 16 min sequence for a three-vehicle class classification problem. Second, the authors achieve a 0·8% counting error rate on 2 h of real traffic.

KEYWORDS: Motorcycle counting, Classification, LIDAR, Road safety, Mobility