Design and Implementation of Intrusion Detection System Based on Monocular Vision and YOLOv5 Algorithm
The fully automatic brick unloading and packaging machine has become a standard configuration for brick and tile building materi-als production lines.During operation,unauthorized entry into the work area by idle personnel can easily lead to safety accidents.The automat-ic alarm upon detecting illegal personnel entering the dangerous area and the automatic shutdown in emergency situations have become the fo-cus of upgrading and renovating the brick unloading and packaging machine.The existing methods for establishing virtual electronic fences based on infrared sensor detection or ultra wideband technology have problems such as low detection accuracy,single warning methods,and difficulty in defining responsibilities after accidents occur.To this end,a fully automatic brick unloading and packaging machine work area il-legal intrusion detection system based on monocular vision and object detection algorithm YOLOv5 has been developed.The system consists of a camera,a microcontroller,an alarm,a relay,and control software.It uses a monocular camera for image acquisition,positioning,and dis-tance measurement,and uses the object detection algorithm YOLOv5 for object detection and recognition;When it is detected that unauthor-ized personnel have entered the work area,corresponding instructions will be sent to the microcontroller,which will be controlled by the sen-sor for alarm,shutdown,and other processing.The simulation experiment results show that the system can effectively complete functions such as photography,positioning and distance measurement,detection and recognition,alarm,emergency shutdown,etc.,with an accuracy of over 94%.Compared with existing methods,it has stronger functions and lower costs,and can effectively solve security problems such as unau-thorized entry of idle personnel in the work area.
万方数据
维普
