Application of computer vision in surveillance cameras to identify criminal approaches
Bruno Cúgola Coelho, André Luís Marques Marcato, Iago Z. Biundini · 2025
Object movement detection through surveillance cameras is increasingly important in areas like public safety, health, and remote monitoring. It helps protect environments and people by identifying suspicious activity and preventing crime. However, there is a need for algorithms that reduce the workload of human operators by enabling simultaneous monitoring of multiple cameras. Implementing YOLOv8 for object detection improves accuracy and efficiency in security systems. Additionally, incorporating fuzzy logic helps address uncertainties and differentiate normal from suspicious behavior. Using Telegram for real-time communication allows sending alerts with detection images. The combination of object identification, reliability filters, and fuzzy logic significantly enhances weapon identification reliability.