Deep learning on rail profiles matching

TL;DR

This paper introduces a deep learning-based method using pre-trained CNNs for accurate matching of rail cross-section profiles, addressing challenges of data volume, shape diversity, and measurement errors in on-site rail profile matching.

Contribution

The paper presents a new high-precision rail profile matching method leveraging deep learning and establishes a large dataset for training and evaluation.

Findings

High accuracy in rail profile matching demonstrated

Deep learning outperforms traditional feature-based methods

Dataset of 46,386 manually matched pairs created

Abstract

Matching the rail cross-section profiles measured on site with the designed profile is a must to evaluate the wear of the rail, which is very important for track maintenance and rail safety. So far, the measured rail profiles to be matched usually have four features, that is, large amount of data, diverse section shapes, hardware made errors, and human experience needs to be introduced to solve the complex situation on site during matching process. However, traditional matching methods based on feature points or feature lines could no longer meet the requirements. To this end, we first establish the rail profiles matching dataset composed of 46386 pairs of professional manual matched data, then propose a general high-precision method for rail profiles matching using pre-trained convolutional neural network (CNN). This new method based on deep learning is promising to be the dominant approach for this issue. Source code is at https://github.com/Kunqi1994/Deep-learning-on-rail-profile-matching.