Advanced image tracking techniques for target seeking

Abstract

The primary goal of visual tracking is to provide a computerized system resembling the capabilities of the human eye. Motion detection is used in many applications and its integration as a facility within a vision system offers many potential benefits. We introduced an investigation of target tracking techniques; two different target tracking algorithms were implemented. The first is Covariance-based target tracking technique and the second is Hu Moments-based target tracking technique. The implemented tracking techniques were investigated and analyzed to determine which algorithm is suitable. Both algorithms were enhanced by achieving segmentation process using the region growing technique and building an adaptive window size that surrounds the target to overcome the problem of scaling. The experimental results show that the enhancement was very efficient not only in target tracking but also in overcoming the occlusion problem.