DocumentCode :
420363
Title :
A volumetric multi-cameras method dedicated to road traffic monitoring
Author :
Douret, Jerome ; Benosman, Ryad
fYear :
2004
fDate :
14-17 June 2004
Firstpage :
442
Lastpage :
446
Abstract :
This paper deals with the issue of using multi-cameras for road traffic monitoring. The aim is to remove the classic monocular ambiguities and to retrieve the objects´ height. An efficient and simple calibration method is introduced. It has the particularity to be connected to the geometry constraints of the road. The method relies on projective geometry and uses the structure of the plane at infinity. In a second stage, a high speed matching procedure is introduced. It is based on an altitude planar decomposition of the road scene. The method naturally achieves two tasks due to altitudes sampling. Match and reconstruction become simultaneous. Finally, experimental results are presented.
Keywords :
calibration; geometry; image matching; image reconstruction; image sampling; monitoring; road traffic; video cameras; altitude planar decomposition; altitudes sampling; calibration method; geometry constraints; image matching; image reconstruction; image sampling; monocular ambiguities; road scene; road traffic monitoring; volumetric multicameras method; Application software; Calibration; Cameras; Computer vision; Geometry; Image processing; Layout; Monitoring; Roads; Vehicles;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent Vehicles Symposium, 2004 IEEE
Print_ISBN :
0-7803-8310-9
Type :
conf
DOI :
10.1109/IVS.2004.1336424
Filename :
1336424
Link To Document :
بازگشت