Improved template update strategy for image alignment

Very few researches about AIA (affine image alignment) algorithm focus on the question of template update strategy which directly affects the performance of the algorithm. In this paper, we firstly analyze the principles of three common strategies and point out their advantages or shortcomings. Then, we propose a new update strategy named FDC (fast drift correction). FDC is to restrain the template drift during tracking by importing a proportion of the first template information as the correction information. However, it makes no compromise with the efficiency of the algorithm. The experimental results show that compared with the traditional update strategies, the update strategy proposed by us performs best currently on both the robustness and the efficiency of the algorithm.