An automated real-time people tracking system based on KLT features detection

The advancement of technology allows video acquisition devices to have a better performance, thereby increasing the number of applications that can effectively utilize digital video. Compared to still images, video sequences provide more information about how objects and scenarios change over time. Tracking humans is of interest for a variety of applications including surveillance, activity monitoring and gate analysis. Many efficient object tracking algorithms have been proposed in literature, however part of those algorithms are semi-automatic requiring human interference. As for the fully automated algorithms, most of them are not applicable to real-time applications. This paper presents a low cost automatic object tracking algorithm suitable for use in real-time video based systems. The novelty of the proposed system is that it uses a simplified version of the Kanade-Lucas-Tomasi (KLT) technique to detect features of both continuous and discontinuous nature. As discontinuous feature selection is subject to noise, and would result in non-optimal feature based object tracking, the authors propose the use of a Kalman filter for the purpose of seeking optimal estimates in tracking. The integrated tracking system is capable of handling shadows and is based on a dynamic background subtraction strategy that minimises errors and quickly adapts to scene changes. Experimental results are provided to demonstrate the system’s capability of accurately tracking objects in real-time applications where scenes are subject to noise particularly resulting from occlusions and sudden illumination variations.