Authors : Kotla chandrika; Yerramasu Krupa Sagar; Silpa Suman; Sowmiya R; Bharani Kumar Depuru

Volume/Issue : Volume 10 - 2025, Issue 3 - March

Google Scholar : https://tinyurl.com/56wumrj2

Scribd : https://tinyurl.com/mb28ebdn

DOI : https://doi.org/10.38124/ijisrt/25mar1270

Google Scholar

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Note : Google Scholar may take 15 to 20 days to display the article.

Abstract : Well-regulated safety is indispensable in scrap-based liquid alloy manufacturing specifically in settings that employ induction furnaces within, in the realm of the metal based scrap industry unit that drives eco-efficient engineering by converting waste into valuable resources strengthening the repurposing of scrap into steel bars reduces dependence on naturally sourced materials enhances resource-efficient energy use and mitigates environmental disruption however the presence of hazardous items such as gas cylinders and pressurized canisters poses significant risks in high-temperature recycling operations To address these challenges we present an automated approach to hazardous substance detection using advanced computer vision techniques our enhanced modern system leverages a custom dataset developed using client- provided and web-sourced images of metal scrap annotated with smart polygon shapes to capture object contours accurately ,where single-shot detector model which is YOLO(You Look Only Once) versions such as yolov9 ,yolov8 and its variants were used and evaluated through extensive data preprocessing and augmentation strategies , yolov9 was selected for deployment due to its superior performance the model achieved a mAP(mean average precision) of 0.86 on test data enabling precise detection and classification of hazardous materials within industrial settings Our solution serves as a safeguard for operational safety, preventing catastrophic events such as chemical reactions, explosions, and toxic emissions that could endanger human lives and disrupt production, as safety becomes important when scraps are melted, as during this process presence of closed substances can cause tremendous effects to environment and workers. Deployed via Streamlit(open- source Python framework), the model provides real-time monitoring of live video feeds, enhancing safety measures and operational efficiency in scrap-based liquid steel production. This automated system not only mitigates risks but also ensures compliance with safety regulations, ultimately improving the integrity and sustainability of industrial processes.

Keywords : Hazard Management, Scrap-Based Liquid Steel Production, Yolov8(You Only Look Once), Computer Vision, Object Detection, Safety, Induction Furnace.

References :

1. Stefan Studer , Thanh Binh Bui  Christian Drescher , Alexander Hanuschkin ,Ludwig Winkler  , Steven Peters  and Klaus-Robert Müller Towards CRISP-ML(Q): A Machine Learning Process Model with Quality Assurance Methodology
2. https://www.mdpi.com/2504-4990/3/2/20
3. Vimal R. Nakum, Kevin M. Vyas, Niraj C. Mehta RESEARCH ON INDUCTION HEATING -A REVIEW
4. https://www.researchgate.net/publication/280313022_RESEARCH_ON_INDUCTION_HEATING_-A_REVIEW
5. Prof. P. D. Kale  COMPARATIVE ANALYSIS OF IMAGE ANNOTATION TOOLS: LABEL IMG, VGG ANNOTATOR, LABEL STUDIO, AND ROBOFLOW https://www.jetir.org/papers/JETIR2405D59.pdf
6. Chien-Yao Wang, I-Hau Yeh2 , and Hong-Yuan Mark Liao YOLOv9: Learning What You Want to Learn Using Programmable Gradient Information https://arxiv.org/pdf/2402.13616
7.  Rejin Varghese, Sambath M.YOLOv8: A Novel Object Detection Algorithm with Enhanced Performance and Robustness

https://ieeexplore.ieee.org/document/10533619

8. Hafedh Mahmoud Zayani , Ikhlass Ammar , Refka Ghodhbani , Taoufik Saidani , Rahma Sellami , Mohamed Kallel , Amjad A. Alsuwaylimi , Kaznah Alshammari , Faheed A. F. Alrslani , Mohammad H.Algarni Unveiling the Potential of YOLOv9 through Comparison with YOLOv8, https://www.ijisae.org/index.php/IJISAE/article/view/5794/4539
9. Jun Li ,Yongqiang Feng,Yanhua Shao,Feng Liu IDP-YOLOV9: Improvement of Object Detection Model in Severe Weather Scenarios from Drone Perspective, https://www.mdpi.com/2076-3417/
10. JeongYoon Rhee, JunHyuk Park, JaeIn Lee,HyunTae Ahn,Long Hoang Pham, JaeWook Jeon A Safety System for Industrial Fields using YOLO Object Detection with Deep Learning https://ieeexplore.ieee.org/document/10210722
11. Ajantha Vijayakumar ,Subramaniyaswamy Vairavasundaram YOLO-based Object Detection Models: A Review and its Applications https://link.springer.com/article/10.1007/s11042-024-18872-y
12. Seong-Eun Ryu and Kyung-Yong Chung
13. Detection Model of Occluded Object Based on YOLO Using Hard-Example Mining and Augmentation Policy Optimization
14. Detection Model of Occluded Object Based on YOLO Using Hard-Example Mining and Augmentation Policy Optimization
15. Kemal Oksuz; Baris Can Cam; Sinan Kalkan; Emre Akbas
16. Imbalance Problems in Object Detection: A Review.