You Only Look Once (YOLOv8) Based Intrusion Detection System for Physical Security and Surveillance

In recent years, the need for advanced physical security and surveillance systems has been increased due to increase in trespassing, theft, vandalism etc. This paper presents an innovative and a novel approach to region-of-interest-based Intrusion Detection System for Physical Security and Surveillance by harnessing the advanced capabilities of You Only Look Once version-8 (YOLOv8), a state-of-the art object detection algorithm, trained and evaluated utilizing the Common Objects in Context (COCO) dataset. This dataset has a comprehensive compilation spanning of 80 diverse object categories which can be used with custom images to train and evaluate the model. Our methodology crafts an efficient and a robust Intrusion Detection System (IDS) for Physical Security and Surveillance employing the Python as a programming language and OpenCV library. Illustrious for its prompt and explicit real-time object recognition, YOLOv8 emerges as the most preferred framework for monitoring systems. Our model takes advantage of the rich information embedded in the COCO dataset to adroitly recognize and categorize intrusions across various scenarios. The study scrupulously delves into the essentials of implementation, encompassing data pre-processing, model training and strategic application of optimization techniques to intensify detection efficiency. The flexible and accessible nature of the Python programming language and OpenCV promotes the creation of a system flawlessly integrable with already existing security frameworks and hardware. Experimental findings emphasize the effectiveness of the proposed region-based Physical Intrusion Detection System in attaining high precision and recall rates for physical intrusion detection. This paper concludes with a subtle discussion, exploring practical implications, potential applications, and avenues for future research within the domain of computer vision-based intrusion detection systems. This research significantly chips in to the evolution of surveillance technology, bolstering security measures across diverse environments.