一种基于主动轮廓模型的自适应模板更新算法

Correlation-based Tracking Algorithm Based on Active Contour Models

在目标相关跟踪过程中,由于目标的姿态、大小发生变化,所以合理地更新模板极为重要。而已有的模板更新方法都不能适应目标的姿态和大小变化。提出了一种模板更新的新方法,该方法以颜色直方图的Bhattacharyya系数为基础,来进行模板更新时机的判断。并利用主动轮廓模型算法得到目标的边缘,自适应调整模板的大小和内容,从而实现对目标在发生姿态和大小等变化下的可靠跟踪。

A new algorithm of correlation tracking algorithm based on active contour models is explored in this paper.Because of the template refresh problem in the traditional correlation,the system with the new technique can track an object which appears in various sharps and perform different scales.The new algorithm considers a color histogram parameter Bhattacharyya as a template refresh rule,and then gets access to the object edge to adaptive refresh template content and scale.