中文摘要：

如何快速准确地实现运动员脚力信息的自动识别对于构建一个智能的、实时的运动员训练指导系统至关重要。文中首次提出了使用支持向量机进行运动员脚力曲线自动分段的方法。在详细讨论和分析抓举运动员的六维脚力曲线特征的基础上，根据曲线对应的五个运动阶段，进行了特征提取，然后构造了基于支持向量机的多类模式分类器，实现了脚力曲线对应运动阶段的自动识别，取得了令人满意的分类效果，为构建举重运动员训练指导系统打下基础。

英文摘要：

How to realize the automatic recognition of a weight lifting athlete＇s ground reaction force（GRF） fast and accurately is very important to the development of training and coaching system. In this paper, we present a new GRF recognition method based on support vector machine （SVM） for the first time. We detail the characteristics of the curves of GRF first and then complete feature selection according to the five motion phases. In the end, we set up a multi-classifier using SVM to recognize the patterns of GRF curves of weight lifting athletes. The experiment results are satisfied. This work lays a solid foundation that develops training and coaching system of weight lifting athletes.