An Object Tracking Algorithm Based on Global SURF Feature

摘要

Aiming at the problem of the object tracking algorithm based on a single feature which can't track the object in the complex cases accurately, this paper proposes an object tracking algorithm based on global SURF feature. Because the color feature is invariant to rotating and deformation, SURF feature has a good robustness to illumination change, a certain degree of angle change, it fuses the two features to from a new feature called global SURF feature. Then, it creates the object model based on global SURF feature, combines with particle filter algorithm and updates the object template and feature points dynamically in the process of tracking. Finally, it realizes the accurate tracking for the object. The experimental results in different settings show that it has a good real-time and high robustness which fits for the object tracking in the complex cases.