Towards Machine Vision Based Railway Assets Predictive Maintenance

Abstract

The main goal of this paper is to present novel technologies that can contribute to safety, competitiveness, efficiency and operational reliability of Railway infrastructure through the development of innovative solutions for measuring and monitoring of railway assets based on machine vision. Measuring the transversal position of the wheels on the rail, as well as identification of the defects of the wheel and the rail (such as deformation of rail head edge, lateral wear, worn wheels, cracks in wheel and rail, rolling contact fatigue, corrugation and other irregularities) can increase reliability and lower maintenance costs. Currently, there is a need on the market for the innovative solution, namely the on-board high-speed stereo camera system augmented with a system that projects custom pattern (fringe scanner system) for measuring the transversal position of the wheels on the rail, robust to environmental conditions and waste along the track that can provide reliable measurements of transversal position of the wheels up to 200 km/h. New trends in Precise Industrial 3D Metrology are showing that stereo vision is an absolute must have in modern specialized optical precision measuring systems for the three-dimensional coordinate measurement.