A Robust On-Road Vehicle Detection and Tracking Method Based on Monocular Vision

Abstract

In this paper, we propose a new framework for vehicle detection and tracking. Multi-features are used in the vehicle detection algorithm, which can be divided into two main steps: generation of candidates using features such as the shadow and vertical edge, and verification of the candidates using HOG and SVM. In the vehicle tracking algorithm, the RGB model and orientation histogram are used to represent the object feature, and the mean shift is employed to search the mode of the potential object rapidly in a neighborhood frame, which obtains the preliminary tracking results. Then, we use ORB feature matching and correction methods to adjust the preliminary tracking results. The improved Mean-Shift tracking results and the ORB correction results are then fused by linear weighted, which obtains the final results of the tracking. Experimental results demonstrate that the proposed approach is robust and validate in complicated real scenes.