摘要
借助计算机视觉技术对运动员的运动动作进行跟踪分析,旨在为运动员提高运动成绩提供科学的技术分析手段.研究序列图像中运动员动作的各特征点的运动变化轨迹,并加以跟踪,提出了基于聚类和α-β-γ滤波的特征点跟踪.为解决聚类数目的不确定性给C均值聚类带来的难点,以α-β-γ滤波预测下一帧图像的特征点的估计值为聚类的中心,从而有效克服了初始聚类中心难以选择的问题.并对40帧的图像序列进行了跟踪实验,结果表明该方法是可行有效的.
The main purpose is to provide a science analysis for the athlete to improve the grade, by computer vision technique to track the movement of the athlete. We proposed a method for tracking the feature points in the image sequence based on clustering and α-β-γ filter. In order to overcome the difficulty caused by the uncertainly of the number in C-mean clustering, the α-β-γ filter was used to predict the clustering center of the moving feature point in the next frame, which effectively solved the difficulty in choosing the initial clustering center. The experiment results demonstrate the practicability and efficiency of the method.
出处
《测试技术学报》
2009年第4期288-292,共5页
Journal of Test and Measurement Technology
基金
国家自然科学基金资助项目(60773204)
浙江省自然科学基金资助项目(Y107124)
