AUTOMATIC VISION BASED INSPECTION OF RAILWAY TRACK: A REVIEW

Ashwini Belkhade, Snehal Kathale

Abstract: Currently, most of railway track inspections are manually conducted by railroad track inspectors. Practically, it is not possible to inspect the thousand of miles of railway track by trained human inspector. This inspection takes too much time to inspect the defected railway track and then inform to the railway authority people. In this way it may lead to disaster. Hence to avoid delay and improve the accuracy, our propose system will automatically inspect the railway track by using vision based method and vibration based method. This method proposes continuous monitoring and assessment of the condition of the rail tracks which prevent major disasters. Our proposed system will inspect the rail track component such as missing bolts, tie plates, anchors etc by using vision based method and simultaneously do the calibration of railway track by using vibration based method. The system provides real-time monitoring and structural condition for railway track using vision based method and calibration to search the fault location on the track. Inspections include detecting defects on tracks, missing bolts, anchor, tie plate and clips etc. In vision based method camera we will use to capture the images or videos. In vibration based method some sensors we will use to detect the vibrations on the railway track. We will extract the signal from 2-D

Keywords: Railway track inspection, Vision based and vibration based method, Image processing, Data acquisition

DOI: https://doi.org/10.15623/ijret.2014.0302003