基于机器视觉的老年人摔倒检测系统

Fall Detection System for the Elderly Based on Machine Vision

家庭陪护机器人可以实现老年人的陪护任务,其中主动检测老年人摔倒情况是一个重要功能,这可以减少独居老年人因摔倒而导致的伤亡。研究基于机器视觉的老年人摔倒检测系统,通过摄像机动态采集场景图像并跟踪场景中的老年人,结合其体姿态特征提取算法,对人体骨骼特征点变化量进行监测,并分析场景的语义信息,自主地对场景中的老年人进行摔倒检测。实验证明,本文提出检测系统是可行有效的。

With the continuous development of economy and the improvement of medical conditions,the birth population is decreasing year by year and the degree of population aging is deepening.The pressure of social pension is great.Facing the lack of pension institutions,home-based pension will be an important trend.Home care robot can achieve the task of care for the elderly,in which it is an important function to actively detect the elderly fall and other abnormal conditions,which can reduce the casualties caused by the fall of the elderly living alone.This paper studies the fall detection system of the elderly based on machine vision,which dynamically tracks the people in the scene through the camera,and combines the human body posture feature extraction algorithm to monitor the changes of human skeleton feature points,and analyzes the semantic information of the scene,and autonomously detects the fall abnormalities of the people in the scene.Experiments show that the proposed method is effective.