Geometric Parameter Identification of Large Bent Pipes Using a Single-View Vision System.

Area of Science:

• Industrial Engineering
• Computer Vision
• Metrology

Background:

• Accurate measurement of large bent pipes is crucial for heavy industry.
• Existing methods for pipe parameter determination can be time-consuming and complex.
• Automated systems for precise pipe metrology are in demand.

Purpose of the Study:

• To develop and describe a vision system for instant measurement of key parameters of large bent pipes (up to 1.2 m diameter).
• To present detailed modeling methods for bending angle, radius, and straight sections.
• To introduce a novel capability for detecting the start and end of these sections in automatic pipe measurement.

Main Methods:

• Utilizing a vision system with a single camera.
• Employing a modified Hough transform for line and curve fitting.
• Implementing image preprocessing techniques.
• Applying distortion models and image projection specific to pipe geometry.

Main Results:

• The system can instantly determine important measurement parameters of large bent pipes.
• Accurate modeling of bending angle, radius, and straight sections is achieved.
• Novel detection of the start and end points of pipe sections is demonstrated.

Conclusions:

• The developed vision system offers an efficient and accurate solution for measuring large industrial bent pipes.
• The system's ability to model and detect specific pipe sections represents a significant advancement in automatic pipe measurement.
• The methods described are applicable to heavy industry settings requiring precise geometric analysis of bent pipes.