摘要
针对传统通过阈值来确定老人跌倒检测算法中的不足和视频检测中容易泄露隐私等缺点,提出了一种基于统计学判决分析的跌倒检测算法。该算法主要通过实验分析来提取行为的特征值并建立特征向量空间,然后利用采样值与行为特征值空间的距离来判断匹配该行为是否为跌倒状态空间的值,同时在算法中提出以智能机器人作为辅助检测的思想。最后通过实验验证了该算法具有较好的鲁棒性。
In view of the shortcoming of the old man fall detection based on threshold algorithm and detection method based video is easy to leak privacy, this paper proposed a decision analysis fall detection algorithm based on statistics. The algorithm extracted the behavior of the characteristic value and established the characteristic vector space mainly through the experiment, then used the sampling values and behavior characteristics of space distance to determine whether the behavior matched the falling value of the state space. At the same time, the algorithm introduced the idea of intelligent robot as auxiliary detection. Finally, through the experiment shows that the algorithm has good robustness.
出处
《计算机应用研究》
CSCD
北大核心
2014年第1期89-91,94,共4页
Application Research of Computers
基金
国家科技部科技型中小企业技术创新基金资助项目(11C26215113536)
关键词
跌倒检测算法
统计学
微电子
传感器
机器人
fall detection algorithm statistics microelectronics sensor robot
