An improved mean shift object tracking algorithm based on ORB feature matching
Yang Yan, Xiaodong Wang, Jiande Wu, Haitang Chen, Zhaoyuan Han · 2015
It is critical to accurately track objects for video monitoring of intelligent transportation, so an improved Mean Shift object tracking algorithm based on Oriented FAST and Rotated BRIEF (ORB) feature matching was proposed in this paper. The algorithm based on ORB feature matching can be applied to better locate the object in case of great shifts to the tracking window when object is interfered by complex background or rapidly moving. Subsequently, the object location can be accurately tracked through Mean Shift iteration tracking. The experimental results suggested that this algorithm had effectively solved following problems, including poor anti-interference performance and inaccurate tracking of fast moving objects. Meanwhile, it improved the robustness of object tracking algorithms.