The proposed Hybrid Edge-AI Surveillance Framework for Real-Time Intrusion Detection of Humans and Animals Using YOLO and LBPH presents an intelligent and automated security solution by integrating artificial intelligence (AI) and Internet of Things (IoT) technologies. The system is implemented on a Raspberry Pi, which is interfaced with a smartphone camera to acquire real-time video streams for continuous monitoring. To enhance detection capabilities, the YOLO (You Only Look Once) algorithm is employed for real-time identification of humans and animals, while the LBPH (Local Binary Pattern Histogram) algorithm is utilized for reliable face recognition of authorized individuals. The system operates by continuously analyzing incoming video frames and identifying potential intrusions. Upon detecting an unauthorized person or suspicious entity, the system captures the corresponding image, stores it locally, and triggers multiple alert mechanisms. These include sending email notifications with captured images, SMS alerts through a GSM module, and activating a buzzer for immediate on-site warning. Additionally, an LCD interface provides real-time system status up-dates. The proposed framework offers a cost-effective, energy-efficient, and scalable surveillance solution capable of real-time monitoring with minimal human intervention. The integration of edge computing with intelligent detection and recognition techniques significantly improves system reliability, reduces false alarms, and enhances overall security for residential and restricted-area applications.
Yadav et al. (Thu,) studied this question.