Plain Line Pattern Recognition, Network Rail, UK

Rail Vision developed the plain line pattern recognition capability for New Measurement Train (NMT) at Network Rail (NR) UK in 2006 which was in use till 2010. As a part of this programme, the linescan data was captured at train speeds of up to 200 kph, to image rail head and rail side/web regions. The train is capable of recording up to 500 miles of data per day. Rail Vision developed the automated image recognition technology for detecting assets and defects within high quality images, and making measurements on size and area of defects. The defects detected included abnormal join gaps, abnormal expansion join width, railhead defects and cracks, fishplate defects, switch area defect recognition, sleeper defects, missing fasteners and weld clamps that should have been removed.

During 2006-2010, Rail Vision provided visual data analysis services to NR for NMT data on West Coast Mainline. Rail Vision is working with the Signalling Team at NR on automated 3D asset mapping , and developing automated risk assessments with sighting distance measurements.