Rail, Turnouts & Switch Inspections
The ability to carry out regular track-level inspections with minimal impact on track velocity is imperative to a safe and efficiently operating railroad network. Our platform enables routine inspection of track and ties (sleepers) using our Machine Learning-powered visual inspection systems. We tightly integrate our platform with our customers’ location referencing systems.
We leverage both GPS and Linear Referencing Systems so that every defect is assigned a unique location on the ground that can be easily referenced by customers using their native systems and languages. Our process is cost-effective, repeatable, and easy to validate, and significantly reduces the amount of track walking required for track assessments and inspections.
How Automated Inspections Work
1. Capture or Supply Imagery
Collected using Cordel sensors or uploaded by the customer.
2. Machine Learning
Analyse data using the Cordel Pixel Classifying ML Engine.
Triage inspection results for desktop review by a trained operator.
4. Produce Reports
Export inspection results into Enterprise Asset Management system.
Imaging System-Agnostic Data Processing
The Cordel platform can ingest all forms of imagery and positioning data, including data from commonly available track imaging systems from well-known providers.
Rail-Native Machine Learning
Our machine learning pixel classifying technology was built entirely in-house alongside railroad engineers. We build and train remarkably powerful rail-specific ML models that improve over time and can be easily modified based on regional and customer specifications.
Powerful and Intuitive Inspection Review
The pixel classifier detects exceptions and queues them for review in our user-friendly and intuitive cloud-based inspection software.
Comprehensive Reporting Based on Engineering Standards
We appreciate that each network has unique engineering standards. So we developed a bespoke process that matches and aligns data outputs from our system to fit each customer's engineering standards.
API Integration with GIS and EAM Software
Once inspections are complete and the human-in-the-loop inspection verification process is completed, the resulting data can then be exported via API, JSON or CSV to our software partners.
Improved Inspection results
A thorough inspection requires the collection, processing, and review of enormous amounts of raw data. By automating the first two steps and using ML to potential flag anomalies, we enable your professionals to spend their time where they're most valuable: making decisions and taking action.
Improved defect ratings
In addition to detecting defects, our machine learning technology can apply estimated condition and severity ratings, enabling us to rank defects by priority when delivering to your technician.
Integrated data outputs and reporting
Data outputs and reports are created automatically delivery engineering-grade reports and alerts via an exception to the relevant operators. Each data layer can be customised to specified engineering standards and requirements