基于改进归一化积相关算法的目标跟踪方法研究

针对归一化积相关算法（Normalizad Product Correlation,NProd）对光照敏感且计算复杂度高的特点,提出一种基于改进归一化积相关的实时目标跟踪方法.通过对NProd算法进行去均值改进,去除图像直流分量,抑制了图像灰度变化对相关峰的影响,提高了算法对光照变化的鲁棒性;通过应用先粗后精的分层搜索算法提高了算法实时性能;最后通过自适应模板更新改善了跟踪算法的稳定性.运动目标跟踪实验表明,该方法保留了NProd算法的高匹配精度,在亮度变化达到25%时可以保证跟踪稳定,且满足实时性要求.

Research on Target Tracking Method Based on an Improved Normalized Product Correlation Algorithm

Normalized product correlation（NProd） algorithm is sensitive for varied illumination and enjoys high computation complexity.Aiming at problems above,this paper proposed an improved NProd algorithm which could suppress the influence of image gray change for correlation peak by removing the DC components of the image,hence the robustness of the algorithm under varying illuminationwas improved.Moreover,by using pyramid searching algorithm and adaptive template update method,the real time performance and stability of the target tracking algorithm were both improved.The results show that the improved NProd algorithm kept the tracking process stable when the variation of illumination reached up to 25%,and the algorithm also satisfied the the real time performance requirement for moving target.