Enhancing Public Safety: Real-Time Weapon Detection through Deep Learning using YOLOv9

Deep learning is an Artificial Intelligence technique which allows the system to learn from the data. One type of artificial neural networks used for image recognition and processing is a Convolutional Neural Network (CNN), which can extract patterns within the images. The challenge is real-time threat identification and response to threats in public spaces such as ATM booths and societies, as traditional surveillance systems depend on human monitoring that can lead to late reactions and oversight. Our solution is a software add-on that scans live CCTV footage in real-time to identify visible weapons. If it locates weapons in the footage, an alarm goes on or a message is sent providing full details of firearms found as described by them to its closest police station. The software works on becoming more accurate and faster. Two different models are trained on two different datasets which produced an accuracy of 92% and 97% respectively. This model is designed to get faster response and eliminate potentially dangerous situations providing a smooth and effective upgrade to existing surveillance systems, which guarantees continued advancement in threat prevention. Bringing an advanced level of security to the masses, this model has a user-friendly interface and can seamlessly mesh with already placed systems. It is simple for clients utilizing coexisting networked encryption facilities.