摘要
为了满足对老年人活动能力的检测需求,提出一种基于人体动作状态序列时序分析的运动模式识别方法。利用加速度传感器采集人体腰部的运动信息,通过滑动窗口对加速度数据进行自动检测、去噪和特征提取,构造隐马尔科夫模型实现人体日常活动序列的训练和识别。实验结果证明该方法可以有效区分不同的日常活动行为,能在辅助医疗中发挥重要作用。
To satisfy the detection demand on activity ability of the elderly, we propose a motion pattern recognition method, which is based on time series analysis of human action states.It utilises the accelerometer to capture the motion information of the waist, and uses sliding window algorithm to automatically detect the acceleration data.After denoising and feature extraction made on the data, it builds hidden Markov model to realise physical daily activity sequence’ s training and recognition.Experimental results show that this method can effectively distinguish between different daily activities, and plays an important role in the adjuvant therapy.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第12期225-228,共4页
Computer Applications and Software
基金
科技部国际合作项目(2010DFA12160)
重庆市科技攻关项目(CSTC
2010AA2055)
关键词
时序分析
运动模式识别
加速度传感器
隐马尔科夫模型
Time series analysis
Human motion pattern recognition
Accelerometer
Hidden
Markov model
