Title :
Enhancement of the road safety with a stereo vision system based on linear cameras
Author :
Burie, J.-C. ; Postaire, J.-G.
Author_Institution :
Centre d´´Autom., Univ. des Sci. et Tech. de Lille Flandres Artois, Villeneuve d´´Ascq, France
Abstract :
In the framework of the Eureka-Prometheus project, the authors have designed an obstacle detection device based on stereo vision. Since raster-scan cameras provide far too much information for real-time processing, linear cameras are used instead and yield an efficient implementation. This paper shows how the linear images shot simultaneously by the left and right cameras are processed in order to detect and locate obstacles in front of the vehicle, how the white lines on the road are used to improve the efficiency and the reliability of the detection, and how a spatio-temporal analysis allows one to study the dynamic behaviour of the detected objects in order to determine the real obstacles. A real-world example is provided
Keywords :
computer vision; feature extraction; image matching; optical tracking; real-time systems; road vehicles; safety systems; stereo image processing; Eureka-Prometheus project; Prolab 2; feature extraction; image matching; linear cameras; linear images; obstacle detection; real-time processing; road safety; spatio-temporal analysis; stereo vision system; tracking; white lines; Cameras; Gunshot detection systems; Image analysis; Object detection; Optical sensors; Road safety; Road vehicles; Stereo vision; Vehicle detection; Vehicle dynamics;
Conference_Titel :
Intelligent Vehicles Symposium, 1996., Proceedings of the 1996 IEEE
Conference_Location :
Tokyo
Print_ISBN :
0-7803-3652-6
DOI :
10.1109/IVS.1996.566369