An Image Processing Based Method For Vehicle Speed Estimation

Vehicle flow estimation is an important part of traffic management system. It plays an important role in tracking systems, automatic video surveillance and also to avoid collision. This paper proposes a method to estimate the speed of vehicles on the highways and city areas. The proposed method can be effectively implemented to control the over speed vehicles and to found guilty in leading to traffic accidents. Each vehicle in the video recorded by the camera is identified. A bounding box is created on the identified vehicle and its centroid coordinates are marked. The analysis of speed is done using mathematical formulae which are embedded in the software. The existing research in this field has certain limitations. The first limitation is consumption of a lot of memory to store videos in the hard drive. The second limitation is inaccuracy of the system in unpleasant weather conditions such fog, haze, rain, and heavy winds, etc. Some systems failed to crate proper bounding box as it is necessary for accurate analysis of the motion of the vehicle and its speed. Another disadvantage is that shadow produced by vehicles on the different lanes of the road creates a fuss and the system detects the shadow too as a different object and creates a bounding box over it. There are other hardware based methods such as radar gun also. Cosine errors occurred when the direction of the vehicle and the radar gun doesn’t match. The objective of the proposed work is to develop a system which can provide the alternative to the radar based systems which can detect multiple vehicles at the same time. We have evaluated the proposed method on various traffic videos and found that the proposed method accurately detect the speed of a vehicle and outperforms many state-of-the-art