Design of Human Detection in Home Security

Electrical Engineering

A home security system is needed to keep objects in the house safe from human interference. This system requires early information on human presence that might endanger security, so the human detection system is one of the most important parts of a home security system. Many people install Closed-Circuit Television (CCTV) at home to maintain home security, so that events that occur around the house can be known. Most of the CCTV installed at home functions as a monitoring and recording of events that occur around the house. With current technological developments, an object such as a human can be recognized and identified through a human object recognition program. By combining a human recognition system with CCTV, as well as a smartphone communication device, a home security system will be produced that can provide information to the smartphone if humans are detected around the house. A human recognition program that is embedded in mini-computer devices such as the Raspberry Pi 4 and gets input in the form of video results from CCTV, then the recorded events around the house can be displayed on a smartphone. In this paper, before the system can recognize human objects, the system first performs the process of detecting human objects using the Single-shot Multibox Detector (SSD) method. The use of the SSD method in detecting human objects has an accuracy rate of more than 80%. By integrating this home security system into a smartphone, users can carry out monitoring functions online and in person and will receive alerts in the form of notifications from applications when video recordings of CCTV detect a human objects.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.