Providing access to expertise in the school of Computing Sciences.

Railway.jpg

Automated Detection of Railway Defects

Business Challenge

Railway tracks must be regularly checked for defects such as loose bolts, broken rail chairs and potentially dangerous debris that could damage assets such as signalling and power equipment. To avoid the time consuming method of a sharp-eyed human walking the track and reporting any problems videos taken from train-mounted cameras are increasingly being used to reduce "walking time". However, those videos are still being checked by a human operator.

SYS Consulting Ltd in collaboration with Snape Signals and Metronet were challenged to develop tools that process the video by automatically searching for specified objects or for changes to the established trackside environment and raising an alert.

Our Solution & Expertise

SYS Consulting has a long record in image processing, with particular strengths in image segmentation, scene understanding and extracting important information from colour and shape within images.

Given that the trackside environment should be almost unchanging over a period of days or even weeks, any changes in the image presented at the same location on successive days would be grounds for raising an alarm. Therefore, SYS Consulting developed an automated computerised monitoring system that efficiently searches each video frame for abnormalities.

Business Benefit

These techniques have the potential to make the operation of a rail network more efficient because track can be monitored frequently without interference with train operation. It also avoids need for manual examination of large amounts of video, or can be used as a double check on manual examination of video or "track walking" examination. Perhaps most important, the system provides a probabilistic output that can be used to prioritize investigation on particular track faults.

More Detail

For further information download the full Automated Detection of Railway Defects Case Study.