مجلة الاربعين المحكمة
Security cameras and video surveillance systems are crucial infrastructures for ensuring the safety and security of the general population. Nevertheless, the identification of high-risk scenarios using these methods is still predominantly carried out manually in several places, particularly in Karbala. Insufficient personnel in the security industry and restricted human capabilities might result in unnoticed dangers or delays in identifying potential threats, so endangering the public during crucial events. As a reaction, many entities have created robotic and automated methods to detect potential dangers by analyzing CCTV footage. The objective of this project is to create an affordable and efficient artificial intelligence (AI) system that can accurately recognize and identify weapons in real-time surveillance films, even in various situations. The system consists of a training phase for the Convolutional Neural Networks (CNNs) using a data set containing images of weapons. The second stage is testing the system, which demonstrated its high ability to detect weapons with high accuracy compared to previous research, the accuracy of the proposed method reached 93%