Steel Wire Rope Surface Defect Detection Based on Segmentation Template and Spatiotemporal Gray Sample Set.

Area of Science:

• Industrial Automation
• Machine Vision
• Non-Destructive Testing

Background:

• Manual inspection of steel wire ropes is labor-intensive and prone to errors.
• Complex environmental factors like lubricants, dust, and lighting variations hinder traditional machine vision defect detection.
• Existing methods struggle with accurate segmentation of wire rope strands and detection of fracture defects.

Purpose of the Study:

• To develop an accurate and robust machine vision method for steel wire rope defect detection.
• To address challenges posed by complex environments and limited defect samples.
• To improve the reliability of automated inspection for critical infrastructure like cableways.

Main Methods:

• A segmentation-template-based method was developed for accurate rope strand segmentation, utilizing structural characteristics.
• A novel dynamic background modeling approach was proposed for fracture defect detection, incorporating spatiotemporal gray sample sets.
• The proposed methods were tested on Z-type double-layer load sealing steel wire rope images from a mine ropeway.

Main Results:

• The segmentation-template method demonstrated high accuracy and insensitivity to light and oil stains.
• The dynamic background modeling approach effectively detected surface defects by constructing a dynamic gray background.
• Comparative analysis showed the proposed method outperformed classic dynamic background modeling techniques (VIBE, KNN, MOG2).

Conclusions:

• The developed machine vision approach offers superior accuracy and effectiveness for steel wire rope defect detection.
• The method exhibits strong adaptability to complex and challenging industrial environments.
• This research provides a more reliable automated solution for ensuring the integrity of steel wire ropes in operational settings.