Robust Technique for Object Tracking by interference of Global Motion Estimation and Kalman Filter

TLDR: A fusion of global motion estimation and Kalman filter-based tracking algorithm is implemented which detects and tracks all the moving objects in the video and achieved a precision of 94.73 percent which is quite good in comparison to other published techniques.

Abstract: In today’s modern world of computer vision there are many techniques for object tracking. But still there is great capacity available for further research. A robust technique for object tracking is proposed in this paper. In this work a fusion of global motion estimation and Kalman filter-based tracking algorithm is implemented which detects and tracks all the moving objects in the video. The algorithm detects corners in a frame and tracks the moving ones in the subsequent frames of the input video. The movement of a moving object is traced by persisting the motion trajectory of corner points on that object. Video stabilization is also implemented so that the moving or shaky video can be processed and amended so that Kalman filter can be implemented. The proposed methodology achieved a precision of 94.73 percent which is quite good in comparison to other published techniques.