Road traffic: Vehicle detection and classification

A good traffic monitoring system should be able to detect, count and classify moving vehicles. Vehicle classification is an important task that can provide information about road users and implicitly take decisions that can reduce congestion for example. This paper presents a new classification approach for moving vehicles. First moving vehicles are detected using a background subtraction approach, then morphological operators are used to reduce the false regions and to remove moving shadows and finally classification is done using the invariant Charlier moments. The performance of the proposed approach is evaluated on real road traffic sequences in comparison of the use of the moments of Hu.