Railway Track Crack Detection System Using Raspberry Pi

Abstract

Railway track monitoring is critical for safe operations but traditional inspection is time-consuming and manpower-intensive. This paper presents RT-MT (Railway Track Monitoring Tool), an automated system using Raspberry Pi with camera, GPS, and GSM modules for real-time track defect detection. Images captured by the camera are processed using YOLOv8 to detect cracks and missing fasteners. When defects are identified, the system sends GPS-tagged alerts to maintenance teams via GSM. Duplicate alert suppression ensures notifications are sent only for new defects at different locations. A web-based platform monitors and records defects in real-time. Evaluation demonstrates 91.3% crack detection accuracy, 87.5% fastener detection accuracy, and 94% location accuracy within 3-meter GPS precision, providing a lowcost automated solution for railway safety.