摘要
本文提出了一种用于鱼眼镜头全方位视觉的组合跟踪算法,该算法将连续自适应均值漂移算法、鱼眼镜头畸变矫正算法和归一化互相关模板匹配算法相结合,解决了目标因完全遮挡或移出镜头视野后再现情况下的目标识别与跟踪问题.实验表明,该方法提高了目标实时跟踪的鲁棒性和精确度并具有较好的实时性.
An advanced combined object-tracking algorithm is presented to implement the real-time tracking of the selected object through omni-directional vision with Fisheye Lens.The new method combined the modified continuously adaptive mean shift algorithm with normalized cross correlation method and fisheye lens correction method.While the object moved out of view or was sheltered completely and then appeared again,this method can detect and track the object quickly.Many good experimental results show that the pro...
出处
《天津理工大学学报》
2010年第2期27-31,共5页
Journal of Tianjin University of Technology
基金
国家科技部"863"计划项目(2007AA04Z229)
国家科技部国际合作项目(2006DHA12410)
天津市自然科学基金(06YFJMC15600)