Copyright © 2025 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{188595,
        author = {Vishwanath and Prof. Ramakrishna Reddy Badveli and Sai Harshavardhan K and E Darshan and Swamy Harthimatt M V},
        title = {HELMET DETECTION SYSTEM},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {7},
        pages = {2313-2318},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=188595},
        abstract = {The inherent hazards and complexity characteristic of dynamic industrial and construction environments necessitate a transition from manual oversight to highly automated, dependable systems for upholding safety compliance, particularly concerning the use of safety helmets. This research presents the methodology for a Helmet Detection System, specifically engineered to provide continuous, real-time safety monitoring. At its core, the solution employs the YOLOv8m deep learning architecture, strategically selected for its capacity to deliver both high precision and rapid inference speed—a vital requirement for surveillance applications. The model underwent specialized training focused on two context-aware classes: person_with_helmet and person_without_helmet. A significant technical contribution of this work is the implementation of a proprietary spatial association logic. This module is critical, as it operates during both the dataset preparation phase and live prediction, confirming the safety status only when a helmet's bounding box is geometrically validated to be positioned within the appropriate upper area of the individual's bounding box. This systematic validation process dramatically enhances the solution's ----robust reliability, effectively preventing erroneous classifications when helmets are merely present in the periphery. The full operational structure is realized through a Python-based Flask back-end responsible for model execution, communicating with a JavaScript-powered front-end that handles video frame encoding and presentation. The final deployment delivers a responsive, instant safety metrics dashboard, demonstrating a practical and superior application of computer vision for enforcing essential safety protocols.},
        keywords = {Object Detection, YOLOv8, Helmet Detection, Safety Compliance, Deep Learning, Context-Aware Vision.},
        month = {December},
        }