Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{186595,
        author = {Mr. Jagdish Rasal and Prof. Nitisha K. Rajgure and Mr. Vineet Shukla and Mr. Jay Thombare and Mr. Shrinath Vathare},
        title = {Intelligent Weapon Detection System For Real Time Survillance Using Deep Learning With YOLOv8},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {6},
        pages = {4654-4658},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=186595},
        abstract = {We present an end-to-end real-time weapon detection system based on YOLOv8 that integrates laptop webcam streams, CCTV footage and still images to detect handheld weapons (pistols, rifles, knives) in varied surveillance settings. The system pipeline includes curated dataset construction (multi-source: webcam captures, CCTV frames, internet/Kaggle), preprocessing/augmentation tailored for small objects, transfer learning with YOLOv8 (variants compared), an alert mechanism (email/SMS/GUI), and an evaluation suite (mAP, precision/recall, F1, FPS latency). We compare lightweight (YOLOv8s) and large (YOLOv8x) variants, run ablation studies on augmentation and super-resolution preprocessing, and demonstrate real-time performance in laptop and edge settings. Our experimental protocol follows prior work on weapon detection and YOLO adaptations in surveillance contexts. (Citations: relevant uploaded works). Results (placeholder) show the proposed pipeline achieves competitive detection accuracy while maintaining real-time throughput — making it practical for deployment in live surveillance scenarios. The system uses over 5000 curated images for training and achieves a mean Average Precision (mAP@0.5) of 86 %, precision of 89 %, recall of 84 %, and real-time throughput of 35 FPS on a laptop GPU. Implemented in Python using Flask, OpenCV, and HTML/CSS/JS, the system features a graphical interface that supports live video detection, alert mechanisms, and logging. Experimental evaluation demonstrates that YOLOv8 delivers superior accuracy and speed compared with traditional CNN-based approaches, making it a practical solution for automated surveillance and public safety enhancement},
        keywords = {YOLOv8, real-time detection, deep learning, computer vision, weapon detection, CCTV, Flask UI, alert system.},
        month = {November},
        }