Real-Time System Based on Feature Extraction for Vehicle Detection and Classification

Today, Road traffic video surveillance becomes the centre of several concerns. It presents an important way for analysis of road traffic in highways. Road traffic video surveillance can help to resolve many problems which can influence road safety. This paper presents a real-time management and control system which serve to analyze road traffic using a stationary camera. The proposed system can measure the quantity and characteristics of traffic in real time based on three modules, segmentation, classification and vehicle counting. Our contribution consists of developing a feature-based counting system for vehicle detection and recognition under the conditions which present a challenge in recent systems, such as occlusions, and illumination conditions. Our method can perform vehicle detection and classification by eliminating the influence of many factors on system efficiency. The obtained results show that the system proposed in this paper provides a counting rate higher than that of some existing methods.