Customer needs 

Detection of missing or damaged slabs, cable ducts, trackside phones, and shelters; open hatches; broken fencing; and landslides near the tracks 

Types of data analyzed 

3D LiDAR images and point clouds

Data acquisition method 

Measurement train

Method of using the digital twin

Creation and training of 3 algorithmic modules for :

• Detection of objects near the railway tracks
• Tracking of objects (Multiple Object Tracking)
• Characterization of changes in the condition of installations

Point cloud segmentation for data analysis and detection of three-dimensional objects

Evaluation of the detection model's performance (proportion of target objects correctly detected by the model)

Benefits for the client

• 1 kilometer of track analyzed in 12 seconds versus X hours on a pedestrian inspection
• 99.5% of detections are accurate
• 7.75 kilometers of track analyzed, with 221 objects detected and classified
• 98.4% accuracy in anomaly detection

What is LiDAR ? 

LiDAR, which stands for "Light Detection and Ranging" or "Laser Imaging Detection and Ranging," is a distance measurement technique based on the analysis of the properties of a light beam returned to its emitter.

A LiDAR system is an optoelectronic device consisting of a laser emitter, a receiver with a light collector (such as a telescope or other optics), and a photodetector that converts light into an electrical signal. It also includes an electronic signal processing chain that extracts the desired information.

LiDAR data collected forms a three-dimensional point cloud with up to 7,000 points per square meter, providing a highly detailed and precise level of information.