Study on Pedestrian Detection Ahead of Vehicle Based on Machine Vision

The pedestrian detection technology has become the main research point in the field of Intelligent Transportation System (ITS) and safety driving assistant. This article proposes a real-time pedestrian detection method. First, the expanded Haar-like characteristic is selected and calculated using integral map. The trained cascaded classifiers can lock the candidate pedestrian areas. Then, the column that has the most distinct edge symmetry through vertical edge extraction can be figured out in the candidate areas. Combined with pedestrian shape features, the width and height of pedestrian can be determined. At Last, the candidate pedestrian is validated based on the gray symmetry and local entropy. The experiment results show that the algorithm is effective and robust.