Vision Based Road Traffic Density Estimation and Vehicle Classification for Stationary and Moving Traffic Scenes during Daytime

Abstract

Automatic road traffic density estimation and vehicle classification are very important aspects of today’s Intelligent Transportation Systems (ITSs). Traditionally loop sensors have been used for this purpose, but lately vision based systems have been preferred due to their advantages and the problems associated with loop sensors. Many vision based vehicle detection and classification algorithms for free flowing traffic have been proposed. These systems are largely dependent on either motion detection or more generally background modelling and subtraction. There is little reported of traffic scenes with very slowly moving or stationary vehicles for which motion detection based approaches are impractical. This paper presents a novel vision based road traffic density estimation and vehicle classification approach that is independent of motion detection and background modelling and subtraction. It combines selected image processing, computer vision and pattern recognition algorithms to obtain the traffic parameters. The approach is applied to both standstill or slow moving traffic, and free flowing traffic under different illumination conditions during the day. The approach does not require camera calibration, therefore, it can work with already installed video surveillance systems, making it economical and convenient. The algorithm is based on image segmentation using a Laplacian of Gaussian edge detector (LoG), morphological filtering of the edge map objects and classification into small, medium and large vehicles on the basis of size using a nearest centroid minimum distance classifier. The proposed approach can be used for both stationary and fast moving traffic in contrast to motion detection based approaches. The algorithm was implemented in MATLAB R2015a and average detection and classification accuracies of 96.0% and 89.4% respectively were achieved for fast moving traffic, while for slow moving traffic, 82.1% and 83.8% respectively were achieved