Reliable Mathematics: NSTU NETI Develops Method for Inspecting Railway Rails with Lidar and Without AI

The work uses fully automated analysis and mathematical algorithms instead of neural networks

Specialists from Novosibirsk State Technical University (NSTU NETI) have developed a method for automated monitoring of the condition of railway tracks. The technology is based on comparing lidar scanning of tracks with a digital reference-"passport" created for sections of any type: straight, turning, and transitional.

The development identifies the slightest deformations of rails caused by wear, temperature changes, or loads, replacing laborious manual inspection. The system accurately localizes problem areas, optimizing the work of repair crews.

The key innovation is a fully automated analysis of lidar data, which is a novelty even at the global level. To calculate the curvature of the path, a combination of mathematical methods is used (difference analogs of derivatives, estimation of bending arrows), which are resistant to measurement errors. Artificial intelligence is deliberately not used at this stage — the emphasis is on reliable mathematical algorithms.

The system improves transportation safety by detecting defects early, reduces the risk of accidents, and optimizes infrastructure maintenance costs. Targeted repairs instead of continuous monitoring are especially relevant for extended Russian highways.

Read materials on the topic:

Russian Railways Implemented AI-Based Video Analytics System to Improve Safety at Transport Facilities

Biometric Boarding for Long-Distance Trains to be Introduced in Russia by 2029

Russian Railways to Implement AI for Managing Tariffs and Train Compositions

Sources
NGTU NETI

Now on home