基于Kinect V2的老人跌倒检测系统设计  

Design of elderly fall detection system based on Kinect V2

作　　者：李沛峰 冯飞龙 黄淑芬[1] 林家豪 LI Peifeng;FENG Feilong;HUANG Shufen;LIN Jiahao(Huali College of Guangdong University of Technology,Guangzhou 510000,China)

机构地区：[1]广东工业大学华立学院,广东广州510000

出　　处：《电子设计工程》2021年第19期112-116,共5页Electronic Design Engineering

基　　金：国家级大学生创新创业训练计划(201813656006)。

摘　　要：随着我国人口老龄化规模的不断增大,养老问题成为社会关注的热点。老年人作为社会的弱势群体,容易发生老人摔倒等问题。如何保障老年人的安全,引发了社会各界关注,也成为社会急切需要解决的问题。为了利用现代化技术对老年人跌倒行为做出准确、快速的检测,提高老人跌倒后的救治率,提出了一种基于Kinect V2图像传感器的老年人跌倒检测系统。该系统通过人体骨骼点的图像识别和骨骼点的跌倒阈值计算行为特征,结合Kinect V2的语音交互系统,能够达到跌倒检测的效果。通过大量的验证表明,该系统能精确、可靠地检测跌倒行为。With the increasing scale of population aging in our country,the issue of pension has become a hot spot of social concern.As a disadvantaged group of society,the elderly are prone to safety problems such as falling down.How to ensure the safety of the elderly has attracted the attention of all sectors of society and has become an urgent problem that needs to be solved in society intensified by aging.In order to use modern technology to accurately and quickly detect the fall behavior of the elderly and improve the treatment rate of the elderly after a fall,the system of elderly fall detection based on the Kinect V2 image sensor is proposed.The system uses the image recognition of human bone points and the fall threshold of the bone points to calculate the behavior characteristics,and the voice interaction system of Kinect V2 can achieve the effect of fall detection.After a lot of verification,the system can be accurately and reliably guaranteed in fall detection.

关 键 词：Kinect V2 跌倒检测 骨骼图像 深度图像 阈值 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]