RAILWAY OBJECTS DETECTION USING IMPROVED YOLO ALGORITHM
Arukala Shreya1, Bellam Vaishnavi2, Kandimalla Likhitha3, Malkedi Navya4, Mrs. M. V. Anjana Devi5, Dr. S. Madhu6
1,2,3,4 UG Scholars,5Associate Professor, 6Professor & Head of the Department
1,2,3,4,6 Department of Artificial Intelligence & Data Science,
1,2,3,4,5,6 Guru Nanak Institutions Technical Campus, Hyderabad, Telangana, India
Abstract - Railway transportation is a cornerstone of modern infrastructure, necessitating robust safety mechanisms. One critical area of concern is the detection of railway objects such as obstacles on tracks, signal indicators, and human intrusions, all of which pose serious threats to railway safety. Traditional object detection algorithms often struggle with the real-time constraints and complex environments present in railway systems. This research proposes an enhanced object detection model based on the YOLOv8(You Only Look Once Version 8) algorithm, optimized specifically for railway applications. The improved algorithm integrates attention mechanisms, advanced data augmentation techniques, and a refined loss function to achieve higher accuracy and faster inference times. The proposed system is evaluated on a curated dataset, demonstrating a substantial performance increase over baseline models. Results show a mean Average Presession (mAP) improvement of 8.5% and a reduction in false positive. The system is suitable for real-time deployment on edge devices, ensuring practical applicability in operational settings. This study contributes significantly to the field of intelligent transportation systems by presenting a scalable and efficient solution for enhancing railway safety through advanced deep learning techniques.
Key Words: YOLOv8, Railway Object Detection, Knowledge Distillation, Deep Learning, Real-Time Detection
