Machine Vision Data Products for the Automated Track Inspection Program [Research Results]

Details

• Creators:
  Baillargeon, Jay ; Stuart, Cameron
• Corporate Creators:
  United States. Federal Railroad Administration. Office of Research, Development, and Technology
• Subject/TRT Terms:
  • Algorithms
  • Flaw detection
  • Image processing
  • Inspection
  • Machine vision
  • Railroad tracks
  • Data
  • Automated Track Inspection Program
  • data products
• Publication/ Report Number:
  RR 21-20
• Resource Type:
  Brief
• Geographical Coverage:
  United States
• Edition:
  Research Results
• Contracting Officer:
  Baillargeon, Jay P;Stuart, Cameron
• Corporate Publisher:
  United States. Department of Transportation. Federal Railroad Administration
• Abstract:
  The Federal Railroad Administration (FRA) Office of Railroad Safety’s Automated Track Inspection Program (ATIP) uses machine vision inspection technology along with Track Geometry Measurement and Gage Restraint Measurement Systems to thoroughly evaluate track. Algorithms trained to detect railroad infrastructure assets (e.g., fasteners) process images collected during ATIP surveys and identify possible defects associated with those assets. With FRA support, ENSCO, Inc. developed software utilities to automatically detect and extract high-resolution images around areas of interest, such as switches, frogs, track geometry defects, and track with reduced lateral strength. Researchers conducted the work from July 2019 to November 2020. These new data products provide additional information for FRA safety inspectors to assess track safety. The first iteration of the algorithms showed promising results that will serve as a basis for the development of an automated evaluation of switch components.
• Format:
  PDF
• Funding:
  693JJ6-19-F-000052
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/57990/dot_57990_DS1.pdf
• File Type:
  [PDF - 809.93 KB]
• Collection(s):
  Federal Railroad Administration
• Main Document Checksum:
  urn:sha256:53891505bc134eb966e15f1d06c72499058cdf04701eefa120c5682f68ca9eb7