Abstract: Detection and tracking of the moving object, people counting and speed measurement are important tasks of many computer vision applications. In this paper, the optical flow system is used for moving object detection. The pyramidal approach of Lucas-Kanade algorithm is used to estimate motion vectors. Because of this technique, there is a reduction in computational burden. Tracking of a moving object is performed using Kalman filter. This technique can track single or multiple moving objects. The number of detected moving objects is counted. People counting useful for the management of the public places. After that, speed measurement is performed. For the traffic management, this technique can be used.

Keywords: Optical flow (O.F.), Lucas-Kanade (L.K.) algorithm, Motion vectors, Kalman filter