Abstract
FRA-mandated railroad track inspections result in significant labor costs and occupy valuable network capacity. These factors, combined with advancements in the field of machine vision, have encouraged transition from human visual inspections to machine-based alternatives. Most commercial systems currently available are limited to a “pass/fail” approach through the detection of components which fail to meet maintenance or safety thresholds, as opposed to being able to detect subtle changes in track conditions to identify evolving problems. To overcome the limitations of human interpretation and simple “pass/fail” defect-finding, this seminar will describe field deployment and validation of a system that pairs three-dimensional (3D) machine vision with Artificial Intelligence and change detection technology. The change detection approach uses a deep convolutional neural network (DCNN) to accurately characterize network conditions between repeat runs before analyzing differences. This seminar discusses the process for 3D image capture and how DCNNs were trained with these 3D images. Finally, it compares the trained DCNNs to an expert human inspector. Preliminary results from this study suggest that this technology can successfully identify present, broken, and missing spikes and fasteners with percent accuracies in excess of 98%.
Speaker
Richard Fox-Ivey has more than 20 years of experience in the Transportation Infrastructure Management field. During this time, he has worked for sensor manufacturers, systems integrators, inspection service providers and asset management system software developers.
In his current role at Railmetrics he is responsible for the company’s activities in the railway industry; working with Railmetrics’ algorithm development team to deliver automated railway inspection solutions for railway owners and operators.
Richard is currently managing his third FRA Research and Development project which involves the development and field-proving of Artificial Intelligence-based railway asset change detection technology.
