基于嵌入式视觉的摔倒警报系统设计与研究  

Design and Research of Falling Alarm System Based on Embedded Vision

作　　者：付可可 刘维来[1] 

机构地区：[1]中国科学技术大学精密机械与精密仪器系,安徽合肥230026

出　　处：《工业控制计算机》2018年第9期72-73,76,共3页Industrial Control Computer

摘　　要：年老体弱者如果在日常生活中摔倒,会受到如骨折等严重伤害,同时也可能是因为如中风等严重疾病发作引起。而此时他们可能已经失去自救能力,如果护理人不能够及时发现并救助,摔倒者病情会迅速恶化甚至危及生命。为此,设计了一种基于嵌入式视觉的摔倒报警系统。首先介绍了该系统的总体设计方案,重点描述了基于Zynq-7000的集图像采集、处理、传输于一体的智能相机的设计。提出了采用最小外接矩形长宽比和人体质心垂直速率变化相结合的算法对摔倒姿态进行识别,仿真证明该算法是可行的。该系统对及时发现年老体弱者跌倒并给予及时救助,提高他们的日常生活安全性具有重要的意义。The overall design of the falling alarm system is introduced,and the design for the camera which has the function of image acquisition,processing and transmission based on Zynq-7000 platform is described emphatically in this paper.As for the analysis of abnormal behaviors such as the fall,the detection algorithm based on the combination of ratio in the vertical velocity of the center and aspect ratio is proposed,the simulation proves that this algorithm is feasible.This system is of great significance for timely detection of old and weak people falling and giving timely assistance to improve their daily life safety.

关 键 词：嵌入式视觉 摔倒警报 ZYNQ-7000 最小外接矩形 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP277[自动化与计算机技术—计算机科学与技术]