A Fine-Grained Method for Detecting Defects of Track Fasteners Using RGB-D Image

To ensure the safety of railway transportation, it is necessary to promote the defect detection of track fasteners. However, most current inspections focus on coarse-grained detection of defects, while neglecting the potential fine-grained defects, such as looseness and subtle deformation. To solve this problem, this paper proposes a method for fine-grained defect detection of track fasteners using RGB-D images. The proposed method is divided into two stages: coarse-grained detection, which detects defects such as missing or loose elastic strips, and fine-grained detection, which detects potential defects such as loose fasteners. By combining coarse-grained detection and fine-grained detection, more excellent defect detection of track fasteners has been achieved The proposed method achieves 99.6% accuracy of coarse-grained detection and 90.6% accuracy of fine-grained detection at 23.3 FPS.