Published Paper Details:

Now a days predicting the motion has become important in our daily lives. Most of the traffic monitoring systems have been using the motion equipment for the estimation of motion of the particular vehicle. Motion Estimation and vehicles segmentation is an interesting solution for the efficient traffic monitoring system. Segmentation of moving vehicles in a scene is very much seen in applications such as robotics, video coding techniques, video surveillance, video indexing. One of the key elements of a traffic monitoring system is the motion analysis component, which segments vehicles from the scene and estimates the motion on the image plane. The objective of the paper is to find the methods that can be used efficiently to extract vehicles from the background in captured video frame and to apply the techniques of motion estimation and segmentation. We have applied Lucas Kanade algorithm to a reference frame and determined the motion vectors of the vehicles. We then extended this by applying segmentation to the reference frame to segment the vehicles from the background and then applied Lucas Kanade algorithm. Lukas Kanade algorithm is selected due to its efficiency in video coding standards and its simplicity for software implementations. This processed information will provide a brief information about the vehicle count, vehicle motion and turning rates at the cross road junction and also in video coding techniques.