摘要： 针对笔尖在书写过程中易受光线、阴影和速度过快等不利因素影响的问题，提出一种基于粒子滤波的多模板笔尖跟踪算法，利用笔尖的形状特点检测定位笔尖，在粒子滤波的基础上融合多模板匹配进行笔尖跟踪。实验结果表明，该算法在采用较少粒子数的条件下，可以避免不利因素的影响，取得较好的跟踪效果。

关键词: 笔尖检测, 笔尖跟踪, 粒子滤波, 模板匹配, 基于视觉的签名

Abstract: For overcoming the lighting, shadow and excessive speed effects during the writing process, this paper proposes a multi-templates pen tip tracking algorithm based on particle filtering. Pen tip is detected according to its shape, and the pen tip tracking is implemented by particle filtering with multi- templates. Experimental results show that this algorithm is effective using fewer particles and the effects of adverse factors can be avoided to some extent.

Key words: pen tip detection, pen tip tracking, particle filtering, template matching, visual-based signature

中图分类号: 

• TP391.4