Design of an automated visual inspection based railway track fault detection system

Abstract

This paper presents a cost-effective, automated system for detecting common railway track defects, addressing the challenges of manual inspection currently practiced in Sri Lanka. The proposed solution leverages advanced machine learning techniques for accurate fault detection, enhancing the efficiency and reliability of railway track maintenance. Sensor fusion is employed to integrate data from multiple sources, while a dynamic digital lighting control system ensures consistent performance across varying environmental conditions. The system is deployed on an unmanned automated rolling stock, minimizing human intervention and improving safety. This innovative approach not only reduces inspection time and labor costs but also enhances the accuracy and reliability of railway maintenance operations, contributing to safer and more efficient railway systems.