The Development of an Automatic Inspection System Used for the Maintenance of Rail Tunnels

Abstract

The inspection of railway tunnels can be time-consuming and costly, relying on manual inspection for the most part that tends to become inefficient with time. The operational environment of tunnels is unaccommodating to the inspecting personnel, by exposing them to dust, humidity, and absence of natural lighting. The purpose of this work is to develop a robotic system to automatically detect for cracks on the tunnel's concrete surfaces. The system will utilize visual inspection methods, using computer vision image processing algorithm for the detection of cracks. Within this work, a concept for a robotic system's prototype, along with developed imaged processing techniques, and conducted experiments used to test and identify the effectiveness and limitations of the vision system's capabilities in crack detection on a mimicked railway tunnel concrete wall. The experimental results have provided comprehensive data to prove the feasibility of such a system for use in automated railway tunnel inspections.