Facial Recognition Robots for Enhanced Safety and Smart Security

摘要

The prominent aspects of technological progress include current technologies such as artificial intelligence (AI), computer vision (CV), and the Internet of Things (IoT) that apply to many sectors, particularly surveillance cameras, which use facial recognition in real time to enhance security. In general, security robots are the most prominent security systems that assist humans in dangerous jobs that humans cannot do. Herein, a low-cost robot prototype running on Raspberry PI 4 model B is proposed for enhanced safety and smart security. This smart robot functions to detect whether an intruder has come inside with the help of IoT and face recognition technology. The broadcasting of photos can take place with the room control through IoT. The system utilizes a facial recognition algorithm, separating authorized persons from intruders. The Multi-task Cascaded Convolutional Neural Network (MTCNN) algorithm is employed in image processing to identify and classify people with an accuracy of 90%. The IoT will notify the room control system if a stranger is detected. An interface is designed to control the robot, which is linked to Wi-Fi from a far end. If the person is recognized as a known person, then authorization is provided to them.