Traditional human-based track inspection is time-consuming, subjective and insensitive to subtle changes in track conditions which may portend a future safety concern. Likewise, traditional sensor-based approaches to inspection, for example geometry cars, are often beyond the budget of Shortline Railroads and are limited to “pass/fail” defect-finding. This webinar will describe the field deployment and validation of a new approach being developed in partnership with the Federal Railroad Administration. This new approach pairs three-dimensional (3D) machine vision with Artificial Intelligence (a deep convolutional neural network) and change detection technology. This webinar will present details relating to the process for 3D image capture on the High Tonnage Loop at TTCI, training the Artificial Intelligence for feature identification, and the detection of changes related to clips, spikes, end of tie ballast levels and joint bar bolting between repeat runs. Attendees will have an opportunity to participate in Q&A at the end of the presentation.
Date: Thursday, October 29, 2020
Time: 2:00 pm - 3:00 pm EDT
Cam Stuart, Program Manager, Office of Research, Development, and Technology, Federal Railroad Administration
J. Riley Edwards, Ph.D., P.E., Sr. Research Scientist and Sr. Lecturer, Rail Transportation and Engineering Center – RailTEC
Richard Fox-Ivey, Principal Consultant, Rail Industry, Railmetrics
FREE for ASLRRA Members
$25 for Non-Members