ON DETECTION AND TRACKING OF THE VEHICLES FROM THE REAL TIME VIDEO STREAM USING BACKGROUND SUBTRACTION PROCESS WITH BLOB TRACKER ALGORITHM

Vehicle detection is an ultimate result of identifying the vehicle as objects and analysis the object parameter like position, total counts and individual speeds to infer the decisions about the smart transportation. There are lots of technology have been formed in the same fashion to detect vehicle on the real time road. Various BS techniques are used to prevail over the issues of illustration variation, shadows, background cutter and camouflage. In this study, a method of tracking and detecting the vehicles from the real time video steaming using a single camera on the road has proposed in which the blob tracker algorithm is used with background subtraction (BS)process to achieve a real high-performance system. The proposed method has real time potentiality and any additional sensor input is not needed to perform this operation. The system’s operation is performed on huge number of still images of vehicles and multiple video scenes in terms of completeness, correctness and overall quality parameters. The performance is evaluated on a data set collected from Bangladesh motorways provided with approximate ground truth gained from laser scans and also evaluated using real time video.