Research and Application on Smart Courtyard Monitoring System

Abstract

Residential safety has always been a basic need for people's lives. In the vast suburbs and villages of China, quantities of courtyard houses are distributed, and the level of security monitoring is related to the people's security and happiness. This study takes the characteristics of the human body with a large aspect ratio into account. By improving the structure of the Faster-RCNN network, a method for enhancing human body features is designed to enhance the network's ability to recognize the human body, and covers various human postures such as standing and lying down, and eventually improves the accuracy of the original human body recognition by 7%. In addition, this study designs and deploys a complete set of real-time monitoring system, which continuously saves images when detecting human intrusion, and sends alarm messages to the user's mobile phone WeChat in real time, which meets the real-time and adds convenience. The actual deployment shows that the system can find the target within 120 milliseconds when the intruder enters the sight and send out alarm messages to meet the realtime design requirements.