跌倒检测系统的研究进展  

The Research Progress of Fall Detection System

作　　者：郑娱 鲍楠[1] 徐礼胜[1,2] 林晓州 黄停 窦元珠 

机构地区：[1]东北大学中荷生物医学与信息工程学院,辽宁沈阳110004 [2]教育部医学影像计算重点实验室,辽宁沈阳110004 [3]熙康阿尔卑斯科技有限公司,辽宁沈阳110179

出　　处：《中国医学物理学杂志》2014年第4期5071-5076,共6页Chinese Journal of Medical Physics

基　　金：国家自然科学基金资助项目(61374015);教育部博士点基金项目(20110042120037);中央高校基本科研业务费探索导向重点项目(N110219001);辽宁省自然科学基金项目(201102-067)

摘　　要：目的:随着社会老龄化的迅速发展,针对老年人群的健康安全监护的需求日益增加。跌倒检测系统能够有效检测老年人跌倒事件的发生并及时报警,保护老年人的健康及安全。方法:对一些常见的跌倒检测的系统、算法、设备及其应用进行了比较研究,并提出了一些关于系统评价标准设定的建议。结果:本文将跌倒检测系统分为基于视频图像的跌倒检测系统、可穿戴式的跌倒检测系统和环境式的跌倒检测系统三大类,并从舒适性、实用性、准确率等方面进行比较,还从实验对象选择、实验环境设置、样本动作分类和检测指标确立四个方面对建立跌倒检测系统的评价标准进行了阐述。结论:本文进行的比较研究系统地概述了跌倒检测系统的常见技术和算法,提出了其评价标准的基础框架,对于其进一步的研究设计、商品化开发具有一定指导意义。Objective： With the rapid development of the aging society, the demand for the health and safety monitoring of the elderly are increasing greatly. Fall detection can effectively detect the fall events of the elderly, which can protect the health and safety of the elderly. Methods： The paper reviews and compares the systems, algorithms, equipment and applications of fall detection, and it gives some suggestions to the evaluation of fall detection systems. Results： The paper divides fall detection systems into three classes, camera-based （or vision-based）, wearable and ambient fall detection systems, and compares them from comfort, practicability and accuracy. It also gives reasonable suggestions for the standard of fall detection systems from four parts： experiment subject selection, experiment action classification, experiment environment building and detecting indicator es tablishment. Conclusions： This paper overviewed the techniques and algorithms of fall detection systems, and presented a simple framework of evaluation, which is important for the further research, design, and commercial development of fall detection system.

关 键 词：跌倒检测 视频图像 穿戴式 环境式 评价标准 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]