Anomaly Detection for Security in Children's Play Areas Based on Image Using Multiple Lines Detection Method

Abstract

This study aims to build a device as a security system to detect anomalies of children moving in play areas based on the Multiple Line Detection (MLD) method in a streaming image. We developed this device to help parents monitor their children's activities when playing in dangerous areas of the home to protect children from kidnapping. In this study, the MLD method can detect the children's activities when playing in three zones: the safe zone with green lines in the image, the caution zone with yellow lines, and the danger zone with red lines. The hardware used to build the devices in this study consists of three components: a camera to stream the image activities of children, a Raspberry Pi to process the image using OpenCV, and a buzzer for early security systems. The results of this study show that when the device detected the children playing in the safe zone, the system commanded the buzzer to turn off. Furthermore, when the camera detects that the children are playing in the caution and danger zone, the device then commands the buzzer to turn on as an early warning security system for the parents.