Blog

Automatic Detection of Objects near and on Railway Tracks

Written by Leslie | May 23, 2024 9:10:03 AM

[Page created in 2024 for translation of original content] 

In the series of articles labeled "R&D," we invite you to delve deeper into the research projects we're developing at Altametris. In this article, discover the research project focusing on the automatic detection of objects near and on railway tracks. 

The Challenge of Automatic Detection of Objects in Railway Environments

The national railway network is extensive, and conducting an inventory of objects of interest along the tracks is an extremely time-consuming task. It requires mobilizing numerous experts to identify various elements near the tracks.

Altametris' Objective

Our AI R&D work involves analyzing railway videos and images (from satellites and photos) to detect and geolocate different categories of objects positioned along several tens of kilometers of tracks. In this way, we aim to automate their inventory.

 

Techniques, Technologies, and Methods Deployed 

We are experimenting with Machine Learning techniques along with various forms of software processing (classical algorithms, mathematical/matrix processes, geospatial intelligence), pushing them to their maximum capacity to provide detection solutions with the highest accuracy.

The image analysis systems resulting from our R&D are positioned on moving vehicles and detect the position of objects to transmit the information to the vehicle operator.

Results and Development Outlook

Whether it concerns the detection of objects near the tracks or the detection of risks and defects on the tracks, our algorithmic technologies have demonstrated favorable results in terms of research, identification, and localization:

  • Telecom objects showing defects (cable trenches, pull boxes, kiosks, service ducts, etc.) near the tracks,
  • Changes to telecom installations near the tracks,
  • The risk of switch point fouling when moving work vehicles on the tracks.

Did you find this article interesting? Do you want to learn more about this project? Do you need assistance in creating your industrial digital twin?