摘要
摘要:在传统的CamShift目标跟踪算法中,仅仅依赖目标的颜色特征,而缺乏目标本身的一些特征,例如,角点,尺度和方向等信息,这将导致跟踪中容易出现中心偏移和无法抵挡目标受干扰等问题。为了能够实时有效地跟踪目标,本文提出一种结合角点特征的CamShift目标跟踪算法。该算法融合了角点的特征不变性,采用基于灰度图像的Harris角点检测算法,在图像的目标区域和候选区域提取包括图像信息的局部特征点,在视频图像相邻帧之间,通过特征点匹配,剔除虚假特征点,得到真实特征点的位置和主方向,用于引导和调整CamShift算法中搜索区域的位置和方向。测试结果表明,与传统的CamShift算法相比,该算法具有更好的跟踪效果和鲁棒性。
In the traditional CamShift target tracking algorithm, it only relies on the color features of the target, but lacks some features of the target itself, such as information of corner point, scale and direction, which will lead to problems such as center deviation and inability to resist target interference. In order to track the target effectively in real time, this paper proposes a CamShift target tracking algorithm based on corner feature. This algorithm combined with the characteristics of the invariance of the corner, and Harris corner detection algorithm based on gray image is used, the image of the target area and candidate region extraction including the local characteristics of image information points, between adjacent frames in video image, through the feature point matching, eliminate the false feature points, get the true position of feature points and the main direction, used to guide and adjust the position and direction of the search area in CamShift algorithm. The test results show that this algorithm has better tracking effect and robustness than the traditional CamShift algorithm.
作者
刘美枝
杨磊
高海
LIU Mei-zhi;YANG Lei;GAO Hai(School of Physics and Electronic Science, Shanxi Datong University, Datong Shanxi, 037009)
出处
《山西大同大学学报（自然科学版）》
2019年第5期14-18,共5页
Journal of Shanxi Datong University(Natural Science Edition)
基金
山西大同大学青年科学研究项目[2018Q1]
