RailAurex
Detect defects from video/LiDAR and prioritize work orders for maintenance of way.
RailAurex is track and catenary inspection software that reads video, thermal and LiDAR data from inspection cars, finds rail, tie, fastener and catenary defects as the car moves, and turns them into geo-tagged work orders in your EAM.
Who is this for?
Maintenance-of-way directors and track engineers at Class I, short line and commuter railroads, and transit agencies with electrified catenary or third rail. It assumes you already run inspection cars or geometry vehicles and an EAM such as Maximo, SAP PM or Infor EAM. It is not a replacement for the inspector's walk; it is the ranking that tells the inspector where to walk first.
What problem does it solve?
A geometry run produces hours of video that nobody has time to watch frame by frame, so the review samples, and sampling is where the hazardous miss lives. The defect that matters is usually on the list already; it just sits below a hundred cosmetic ones with the same tag.
Findings then have to be typed into the EAM by hand, with a milepost someone estimated from a screenshot. By the time a crew is scheduled the list is a month old, and on a short line with three people and two hundred miles the list never gets fully walked at all.
How does it work?
The models run on an edge device in the inspection car and classify defects during the traversal itself: rail cracks, gauge anomalies, fastener defects, tie degradation and catenary wear, from visible, thermal and LiDAR inputs fused together. No connectivity is needed on the route; the findings sync to your GIS and EAM when the car is back in range.
Every defect carries GPS coordinates and a milepost reference, overlays on ArcGIS, QGIS or your own GIS, and can be filtered by type, severity and date. Severity is scored as cosmetic, minor, major or critical, weighted by location (mainline, siding, yard) and, if you want, by traffic volume, speed restrictions and maintenance windows. The score orders the work orders that go to Maximo, SAP PM or Infor EAM, with two-way sync on status.
Accuracy is checked, not assumed. Inspectors verify a sample of findings through QA playbooks, and precision, recall and F1 are tracked by defect type and route on dashboards that also watch for model drift. On tuned corridors detection sits above 95 percent; on a new corridor it starts lower, and we show you the number while it climbs.
Capabilities
- Video/LiDAR ingestion; defect detectors
- Geo-tagged findings; GIS overlay
- Work order export to EAM
- Model accuracy dashboards
- Route-based inspection planning
Outcomes
- Inspection time ↓ 30–50%
- Defect detection ↑ 95%+
- Hazardous misses ↓ Zero
What does it connect to?
- EAM and CMMS: IBM Maximo, SAP PM, Infor EAM and Oracle EAM through REST APIs, with two-way work order status
- GIS: Esri ArcGIS, QGIS or a custom platform, with milepost referencing
- Inspection car sensors: visible-spectrum and thermal cameras, LiDAR
- FRA reporting integration for Class I and passenger operators
How is it deployed, and how long does it take?
Edge devices ride in the inspection car and run offline during the traversal; a cloud or on-prem server receives the findings for the GIS overlay and the EAM hand-off. Short lines usually take the cloud tier for cost; Class I and transit operators tend to keep processing on-prem.
Night-run inspections suit commuter and transit networks with tight maintenance windows, since the sensors do not depend on daylight.
GIS overlay and EAM integration live in 8 to 12 weeks, with inspection time typically down 30 to 50 percent once the ranking is trusted.
Industry context
Explore the related workflows and other solutions for your industry.