Moving object detection from image sequence in context with multimedia processing

Research in motion analysis has evolved over the years as a challenging field. Moving object detection and tracking is of major interest in applications such as traffic monitoring, military, medicine and biological sciences etc. In this a new algorithm is proposed for the detection and tracking of moving objects. The image sequence is taken from the digital camera. After preprocessing, the first two frames from the image sequence are segmented by using 8 component connectivity. These segments/clusters are used as seeds for the next frame segmentation. From these segments centroid and area of each cluster is calculated. The Kalman estimation is used to estimate the centroid values for the next frame. The moving object is detected and tracked by using Kalman estimated centroid values. Estimated centroid values of the current frame and previous frames are compared and if there is enough change in the centroid position, it means it is the cluster which belongs to the moving object.