Real time rail line anchor inspection : context of Bangladesh

Abstract

Rail inspection is an essential task in railway maintenance. It is periodically needed for preventing dangerous situations and ensuring safety in railways. In Bangladesh it has been seen many train accidents occur due to missing of hooks which attach the tracks to the ground. At present, this task is operated manually by a human operator who periodically walks along the track searching for visual anomalies. This manual inspection is lengthy, laborious and subjective. This thesis presents a machine vision-based technique to automatically detect the presence of rail line hook using Shi-Tomasi and Harris-Stephen feature detection algorithms collaboratively. This inspection system has been carried out on videos acquired by a digital camera installed on a cart. The combination of these two algorithms has been successful to identify scenarios with attached and missing hooks with accuracy of 85.7%. Hence it can be concluded that the proposed system is robust.