| | |
Summary: Multispectral Machine Vision for Improved Undercarriage
Inspection of Railroad Rolling Stock
Benjamin Freid, Christopher P.L. Barkan, Narendra Ahuja*,
John M. Hart*, Sinisa Todorvic*, Nicholas Kocher
Railroad Engineering Program - Department of Civil and Environmental Engineering
*Computer Vision and Robotics Laboratory - Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
Summary: Multispectral machine vision systems have the potential to provide mechanical inspection
personnel with a tool to automatically assess and monitor the condition of rolling stock. By incorporating
the visible and infrared spectra with machine vision algorithms, such a system can analyze physical and
thermal condition. If a machine vision algorithm determines that a component is outside its normal
operating range, the other spectrum can be analyzed to determine if there is any correlated anomaly. We
present preliminary results on a project developing multispectral machine vision technology for
inspection of the undercarriage of rolling stock. We have developed a system to record digital video
images from a below-track perspective, and several machine vision algorithms to identify and analyze
features of interest in these images.
Index Terms: automated mechanical equipment inspection, machine vision, infrared, condition
monitoring, safety, electrical, component
1. INTRODUCTION
Current practices for inspection of railroad rolling
|
