Zhuojin Pipe Fitting Co., LtdQt-Based Online Wall Thickness Detection Software Design for Steel Pipes
Literature Overview
This 2023 study, published in the journal Machine Tool and Hydraulics (Vol. 51, No. 7, pp. 93–99) by Liu Kun, Tu Deyu, Zhu Qing, and Liu Qingyun from Anhui University of Technology, presents the design and implementation of an online wall thickness detection system for steel pipes based on machine vision technology. Funded by the Anhui Provincial Natural Science Research Key Project (KJ2021A0403) and the Anhui Provincial Science and Technology Major Special Project (201903a05020029), the research addresses the practical challenge of replacing manual sampling inspection with automated, continuous, and accurate wall thickness measurement during steel pipe production.
Core Technical Findings
System Architecture and Machine Vision Approach
The system is designed to perform online wall thickness measurement of steel pipe end faces using machine vision technology. The approach involves camera calibration to establish the relationship between image pixel values and actual physical dimensions, followed by image acquisition, preprocessing, edge feature extraction, and circular contour detection. The system achieves a measurement accuracy of approximately 0.1 mm, which is a significant improvement over manual sampling inspection in terms of both accuracy and efficiency.
Improved RHT Circle Detection Algorithm
A key technical contribution of this study is the improvement of the Randomized Hough Transform (RHT) circle detection algorithm for detecting the circular contours of steel pipe end faces. The conventional Hough Transform (HT) is computationally expensive for real-time applications, while the RHT reduces computational cost but may sacrifice accuracy. The improved RHT algorithm balances computational efficiency and detection accuracy, enabling real-time online detection during production.
| Component | Description | Performance |
|---|---|---|
| Camera calibration | Establishes pixel-to-physical dimension mapping | Enables accurate dimensional measurement |
| Image preprocessing | Noise reduction, contrast enhancement | Improves edge detection quality |
| Edge feature extraction | Identifies inner and outer edge points | Foundation for circle fitting |
| Improved RHT algorithm | Detects circular contours of pipe end face | Real-time capable |
| Wall thickness calculation | Computes difference between outer and inner radii | Accuracy approximately 0.1 mm |
| Qt-based interface | Cross-platform human-machine interaction | User-friendly operation |
Qt-Based Human-Machine Interaction Interface
The system interface is developed using the Qt cross-platform development framework, which enables the software to run on different operating systems without modification. The interface provides operators with real-time visualization of the detection process, measurement results, and quality status indicators. The design emphasizes ease of use, allowing operators to quickly configure detection parameters, review measurement data, and identify out-of-specification pipes.
Engineering Practice Integration
Comparison with Traditional Inspection Methods
| Method | Efficiency | Accuracy | Data Volume | Cost |
|---|---|---|---|---|
| Manual sampling inspection | Low | Low | Small | Low |
| Ultrasonic testing (UT) | Medium | High | Medium | Medium |
| Machine vision (this study) | High | High (0.1 mm) | Large | Medium |
| Laser scanning | High | Very high | Large | High |
Implementation Considerations
For engineers considering implementing similar systems in steel pipe manufacturing facilities, several practical considerations arise:
- Lighting conditions: Machine vision systems are sensitive to lighting conditions. The lighting setup must be carefully designed to ensure consistent illumination of the pipe end face, minimizing reflections and shadows that could affect edge detection.
- Pipe surface quality: The accuracy of wall thickness measurement depends on the quality of the pipe end face. Rough cuts, burrs, or oxidation can affect edge detection. The system may need to be integrated with end-face preparation processes.
- Throughput requirements: The detection speed must match the production line speed. The improved RHT algorithm is specifically designed for real-time performance, but the actual throughput depends on the camera frame rate, processing speed, and pipe speed.
- Environmental factors: Steel pipe production environments may involve dust, fumes, and vibration, which can affect camera performance and image quality. The system design must account for these environmental challenges.
- Data management: The system generates large volumes of measurement data. A robust data management system is needed for storage, analysis, and quality traceability.
Quality Control Integration
The online detection system can be integrated into a comprehensive quality control framework:
- Real-time feedback: Immediate identification of out-of-specification pipes allows for immediate corrective action.
- Statistical process control (SPC): The large volume of measurement data enables SPC analysis to identify process trends and potential quality issues.
- Traceability: Each pipe can be associated with its measurement data, enabling full traceability from production to delivery.
- Process optimization: Analysis of wall thickness distribution can guide process parameter optimization to improve dimensional consistency.
Key Questions and Reflections
- How does the measurement accuracy vary with pipe diameter and wall thickness? The study reports 0.1 mm accuracy, but this may not be uniform across all specifications.
- What is the system's performance on pipes with non-circular end faces (ovality, out-of-roundness)?
- How does the system handle pipes with coatings or surface treatments that affect image contrast?
- What is the long-term reliability and maintenance requirement of the system in a production environment?
- Can the system be extended to detect other defects, such as end-face flatness, chamfer quality, or surface cracks?
Study Insights and Implications
This research demonstrates the practical application of machine vision technology for online quality control in steel pipe manufacturing. The improved RHT circle detection algorithm represents a meaningful technical contribution that balances computational efficiency and detection accuracy. The Qt-based interface design ensures cross-platform compatibility and user-friendly operation, which are important for industrial deployment. For steel pipe manufacturers, this type of system offers significant benefits in terms of inspection efficiency, measurement accuracy, and data availability for quality improvement. The 0.1 mm measurement accuracy is sufficient for most steel pipe applications, and the continuous inspection capability provides comprehensive quality data that can drive process optimization. As steel pipe production continues to demand higher quality and tighter tolerances, online machine vision inspection systems will become increasingly important components of modern manufacturing infrastructure.