Railway-Fastener Point Cloud Segmentation and Damage Quantification Based on Deep Learning and Synthetic Data Augmentation

Abstract

Accurate detection and quantification of damage to railway fasteners are crucial for ensuring railway safety. The spatial damage defects caused by the complex shape of fasteners and the problem of data imbalance in actual scenarios are significant challenges faced by deep learning models. This study innovatively proposes a railway-fastener point cloud analysis method based on deep learning as follows: (1) use four cameras to capture three-dimensional point cloud data and construct a virtual negative sample supplementary data set, (2) develop Rail-Swin3D models for precise segmentation of fastener components, and (3) introduce quantitative indicators to objectively evaluate the damage situation. A data set containing 120 real and virtual damaged fasteners was ultimately constructed, achieving up to 99.35% mean intersection over union (mIoU) in point cloud segmentation tasks. This study not only improves the efficiency of railway safety detection, but also opens new paths for the application of point cloud data in the field of railway maintenance, with profound theoretical and practical value. Accurate detection and quantification of damage to railway fasteners play a crucial role in ensuring railway safety. This research introduces an innovative approach that leverages deep learning techniques and railway point cloud data to address challenges related to spatial defects, data imbalance, and quantitative assessment of fasteners. By using this method, highly accurate point cloud segmentation of railway fasteners can be achieved, followed by the objective evaluation of three-dimensional damage. Our work applies point cloud data to railway maintenance, enhancing the precision and efficiency of damaged railway fasteners detection.