The identification of traffic flow based on edge symmetry

Abstract

Traffic in the megacity has characteristics of high vehicle density, small space between vehicles, which results in a greater distress to the vehicle identification and statistics. This paper designs and implements a vehicle identification algorithm based on edge detection and symmetry. The algorithm can meet the basic needs of urban traffic statistics, and overcome the influence of external factors such as inclement weather. At the end of the paper, some experiments are taken to verify that this method made a great classification accuracy of 94% under good weather and 86% at rainy day.