Title :
Discrimination of an Approaching Vehicle at an Intersection Using a Monocular Camera
Author :
Eguchi, Junji ; Koike, Hiroyuki
Author_Institution :
Honda R&D Co., Ltd., Tochigi
Abstract :
A method of discriminating an approaching vehicle was devised for the purpose of constructing a visual driver assistance system at an intersection using a monocular camera. When a monocular camera is used, it is basically difficult to acquire distance information, so vehicle detection and discrimination of approaching vehicles without the use of distance information are associated issues. In this research, the presence of a vehicle at an intersection is detected by classifying the region on the image that has the shape of the vehicle, and the degree of the vehicle approach is discriminated by comparing the motion vector of the vehicle with the subject vehicle´s direction of movement. By performing a test using this method, it was confirmed that it is possible to discriminate approaching vehicles.
Keywords :
cameras; driver information systems; image classification; image sequences; object detection; road accidents; road traffic; road vehicles; image classification; monocular camera; motion vector; optical flow; road accident; road traffic; vehicle detection; vehicle discrimination; visual driver assistance system; Intelligent vehicles; Motion detection; Performance evaluation; Research and development; Road accidents; Shape; Smart cameras; Testing; Vehicle detection; Vehicle driving;
Conference_Titel :
Intelligent Vehicles Symposium, 2007 IEEE
Conference_Location :
Istanbul
Print_ISBN :
1-4244-1067-3
Electronic_ISBN :
1931-0587
DOI :
10.1109/IVS.2007.4290184