Vehicle Detection by means of Stereo Vision-based Obstacles Features Extraction and Monocular Pattern Analysis

This paper presents a stereo vision system for the detection and distance computation of a preceding vehicle. It is divided in two major steps. Initially, a stereo vision based algorithm is used to recover the extract relevant 3D features in the scene, these features are furtherly investigated in order to select the ones that belong to \emph{vertical} objects only and not to the road or background. These 3D vertical features are then used as a starting point for preceding vehicle detection; thanks to a symmetry operator a match against a simplified model of a rear vehicle's shape is performed using a monocular vision-based approach that allows the identification of a preceding vehicle. In addition, thanks to the 3D information previously extracted an accurate distance computation is performed. This system has been developed in the framework of a cooperation between the Università di Parma (Italy) and the I.N.S.A. of Rouen (France).