Obstacle detection in a sequence of rearview images

Author

Yang, Rong ; Cabestaing, F. ; Postaire, Jack-Gérard

Author_Institution

Centre d´´Autom., Univ. des Sci. et Technol. de Lille, Villeneuve d´´Ascq, France

Volume

1

fYear

1997

fDate

9-12 Sep 1997

Firstpage

200

Abstract

We present an obstacle detection framework which aims at avoiding collisions during overtaking on a highway. Potential obstacles are detected by a vision system which processes the images given by a video camera installed near the left lateral rearview mirror of the car. Images of the sequence are resampled, following an adaptive model of the road geometry. Then, a motion analysis step allows an efficient discrimination between dangerous and non-dangerous potential obstacles. Motion estimation is carried out by a time delayed neural network (TDNN)

Keywords

adaptive signal processing; automobiles; computer vision; delays; image sampling; image sequences; motion estimation; neural nets; video cameras; video signal processing; TDNN; adaptive model; car; collision avoidance; highway; image resampling; left lateral rearview mirror; motion analysis; motion estimation; obstacle detection; rearview image sequence; road geometry; time delayed neural network; video camera; vision system; Cameras; Delay effects; Geometry; Machine vision; Mirrors; Motion analysis; Motion estimation; Road accidents; Road transportation; Solid modeling;

fLanguage

English

Publisher

ieee

Conference_Titel

Information, Communications and Signal Processing, 1997. ICICS., Proceedings of 1997 International Conference on

Print_ISBN

0-7803-3676-3

Type

conf

DOI

10.1109/ICICS.1997.647087

Filename

647087