LI Jinghui, CHI Zongtao, LI Zhongxiao. Human fall detection system based on threshold analysis method[J]. 2019, (8): 80-82+86. DOI: 10.13873/J.1000-9787(2019)08-0080-03.
基于阈值分析法的人体跌倒检测系统
摘要
为了在老年人跌倒时及时发现并进行救助
设计了一个跌倒检测、定位报警求救的智能系统。系统使用STM32F103ZET6作为微处理器
采用ADXL345三轴加速度传感器采集数据
并用ATK—NEO—6M实现全球定位
跌倒发生时
SIM900A发送包含位置信息的求救短信给特定人员。提出了一种跌倒检测算法
根据设定的合加速度阈值和时间阈值
来检测失重、撞击、静止三个过程是否顺序发生
根据设定的姿态角阈值
检测人体姿态角是否超过正常范围
从而判断是否跌倒。测试结果表明:系统性能稳定
检测跌倒的准确率达到97%
满足人体跌倒检测的标准。
Abstract
In order to find and rescue in time when the elderly fall
an intelligent system for fall detection and positioning alarm and resuce is designed. STM32 F103 ZET6 is used as the microprocessor of the system
triaxial acceleration sensor ADXL345 is used to obtain acceleration data. And ATK—NEO—6 M is used to realize global positioning. When the elderly fall
SIM900 A sends a message containing location information for help to a specific person. An algorithm for fall detection is proposed. The resultant acceleration threshold and time threshold are used to determine whether the three processes of weightlessness
impact and rest occurred sequentially. The human attitude angle is judged by the set attitude angle threshold whether it exceeds normal range
and the system determines whether or not the person fell eventually. Test results show that the accuracy of fall detection is 97 %.The system has stable performance and meets the standard of human fall detection.