In target tracking problems, we need to obtain the tracking information of the target with high accuracy and process it in real time. For tracking a target in image frame sequence, we should consider a proper target tracking algorithm.
Generally, the centroid estimation algorithm is widely used to obtain the tracking information of the target. To estimate the centroid of the target, we need to segment the target region from noisy background by using threshold values in current image frame. But we cannot obtain the tracking information of the target with high accuracy by simple or heuristic thresholding methods . Thus, we need to optimize or at least sub-optimize the threshold values for the target region segmentation with more accuracy.
In this paper, various algorithms that have been used are surveyed and their limitations are discussed. Then, we proposed a new adaptive algorithm which sub-optimizes the threshold values for image segmentation in each image frame. The simulation results are given to verify the validity and advantages of the pro-posed algorithm.