A Novel Framework for Detecting Slow Moving Vehicles in Urban Traffic Using Robust Visual Features

Abstract

In the era of Intelligent transportation systems, traffic congestion analysis based on moving vehicles is gaining tremendous importance due to its significance as well as complex intrinsic ingredients. However, in the existing literature, traffic congestion analysis for urban roads are mainly carried out based on vehicle speed, which may drastically vary depending upon the type of vehicle. On the other hand, in the current urban traffic scenarios, one of the main factor for heavily congested traffic is, the presence of slow moving vehicles(such as Bulldozers and Excavators), which occupy a huge portion of road and also fail to coop with average speed of other moving vehicles. In this way, the slow moving vehicles can affect the traffic flow characteristics to a greater extent, due to which their detection as well as tracking in traffic-prone roads are very much essential in order to design efficient traffic facilities in urban regions. In this paper, a new visual features based framework is presented, which can detect the presence of slow moving vehicles by means of employing SURF-based features. The experimental results carried out on training as well as testing datasets demonstrate the reasonable performance of the proposed framework, which can be further incorporated in congestion evaluation systems.