基于图像的老年人马路意外摔倒昏厥自动监测系统设计

The Design of Automatic Monitoring System on People Fainting on Road

针对现有道路状况监测系统对老年人摔倒等意外缺少必要的监控功能,设计了一套基于计算机分布式视觉模型的可监控行人意外状况的视觉监控系统。监控系统设计了底层控制站视觉信息采集模块和DP总线信息传递模块,准确、实时地检测出摔倒、昏厥、长期倒地等针对老年人的意外事件,增加了当前监控系统功能。系统利用前100帧道路实时图像的均值作为初始背景;根据不同意外事件获取背景图像进行实时更新;利用差分法获取差分图像,并进行形态学处理;根据DSP处理器得到的老年人不同姿态的二值图上运动目标的特征,判断发生了哪类意外事件。针对该系统的实验结果表明,系统对摔倒、昏厥、行走等事件检测正确率达到86.5%,适用于信息融合型的交通智能监控系统应用。

In view of the existing road condition monitoring system for tile elderly fall accident lack tile necessa- ry monitoring functions such as, design a set of model based on distributed computer vision visual monitoring system ~an monitor the pedestrian accident situation. The underlying control station monitoring system design visual infor- mation collection modules and DP bus information transmission module, accurate, real-time detect fall, fainting, long-term fall on the ground for the elderly, such as accidents, increase the function of the current monitoring sys- tem. System takes advantage of the average 100 frames before road real-time image as the initial background; ac- cording to the dift^rent accident to obtain real-time update the background image; hy using finite difl＇erence method for difference image and morphological processing; based on DSP processor in different posture of binary chart of ehterly people, the characteristics of the moving target, judge what kind of accident happened. The experimental results show that the system for the system to fall, fainting, walking event detection accuracy reached 86. 5% , amt is suitable for the type infi）rmation fusion of intelligent traffic monitoring system application.