Real-Time Edge-Based Burglary Detection and Automated Alerting Using Deep Learning Framework

Abstract

Burglary stands out as the major concern in terms of security, especially in India's major cities, where the National Crime Records Bureau reports over 100,000 burglaries. This paper proposes an edge-intelligent system that could potentially deter burglars using a real-time CNN-based deep learning model. The system is divided into two stages. Stage 1 uses a face recognition model to identify authorized individuals; hence, no action is taken if a recognized person is identified. Stage 2 is initiated; this stage detects burglary action using a burglary action detection model. This model is deployed on the edge device to prevent cloud security threats. Once a high confidence level is received indicating a burglary action, the system sends an SMS notification and a video feed using the Twilio REST API. Experimental results show the reliability of the system for real-time residential applications.