Object Tracking Based on Meanshift and Particle-Kalman Filter Algorithm with Multi Features
Irene Anindaputri Iswanto, Tan William Choa, Bin Li · Procedia Computer Science · 2019
Object tracking is considered to be a key and important task in intelligent video surveillance system. Numerous algorithms were developed for the purpose of tracking, e.g. Kalman Filter, particle-filter, and Meanshift. However, utilizing only one of these algorithms is considered inefficient because all single algorithms have their limitations. We proposed an improved algorithm which combines these three traditional algorithms to cover each algorithms drawbacks. Moreover we also utilized a combination of two features which are color histogram and texture to increase the accuracy. Results show that the method proposed in this paper is robust to cope with numerous issues, e.g. illumination variation, object deformation, non linear movement, similar color interference, and occlusion. Furthermore, our proposed algorithm show better results compare to other comparator algorithms.