Parametric Reconstruction of Tee Pipe Models via Overlapping Point Cloud Registration

Literature Overview

This paper by Jiang Linjun, Huang Yunbao, Song Zikang, Li Haiyan, and He Yong, published in Modular Machine Tool and Automatic Manufacturing Technique (2026, No. 7), presents a parametric reconstruction algorithm for tee pipe models based on overlapping point cloud registration. The study addresses the challenges of low efficiency, unstable quality, and poor multi-view point cloud registration accuracy in welding robot applications. The research was supported by the National Natural Science Foundation of China (Project No. 12127810). This topic is directly relevant to the field of pipe fitting welding, where accurate model reconstruction is essential for robotic welding quality.

Algorithm and Methodology

The proposed algorithm consists of several key steps. First, multi-view point clouds are acquired through laser camera scanning of the tee pipe and preprocessed through coordinate system transformation to achieve a good initial pose. Second, the Intrinsic Shape Signature (ISS) algorithm and normal vector angle threshold are used to extract key points, and a local angles and curvature histogram (LACH) feature descriptor is constructed for matching to extract overlapping point clouds. Third, a point-to-plane ICP (Iterative Closest Point) algorithm with linear least squares optimization is employed for precise registration, achieving point cloud stitching. Finally, the complete stitched point cloud is registered with the discrete point cloud of the workpiece CAD model for secondary registration, completing the tee pipe model reconstruction and extracting the weld seam for robotic welding.

The experimental results demonstrated that the proposed algorithm improved point cloud registration accuracy by 65.46% compared to other algorithms, and the average error of the extracted weld seam was less than 1 mm, meeting welding process requirements. The following table summarizes the algorithm performance:

Engineering Application to Welding

The application of this algorithm to tee pipe welding has significant practical implications. In pipe fitting fabrication, tee pipes are commonly welded to main pipelines, and the accuracy of weld seam location directly affects welding quality. Traditional methods of weld seam location rely on manual measurement or simple coordinate transformation, which are time-consuming and prone to errors. The proposed algorithm offers a rapid and accurate alternative that can significantly improve welding efficiency and quality consistency.

The use of the LACH feature descriptor is particularly noteworthy. In the context of pipe fitting welding, the geometric features of tee pipes, including the curvature at the branch intersection and the angles between the branch and run, are critical for accurate registration. The LACH descriptor captures both angular and curvature information, making it well-suited for the complex geometry of tee pipes. This is analogous to the feature extraction methods used in welding defect detection, where geometric features are used to identify and classify defects.

The point-to-plane ICP algorithm with linear least squares optimization offers improved convergence and accuracy compared to traditional point-to-point ICP. In welding applications, where the accuracy of weld seam location is critical for achieving proper weld geometry and penetration, this improvement is significant. The requirement that the weld seam extraction error be less than 1 mm is consistent with the tolerances specified in welding standards such as ASME B16.9 and GB/T 12459 for butt-weld fitting fabrication.

Reflections on Welding Automation and Quality

This study represents an important advancement in the automation of pipe fitting welding. As the demand for high-quality, consistent welds increases in industries such as petrochemicals, power generation, and nuclear engineering, the ability to accurately locate weld seams and execute automated welding becomes increasingly important. The proposed algorithm addresses a key challenge in this process: the accurate reconstruction of complex pipe geometries from multi-view point cloud data.

The integration of machine vision with welding robotics offers a path toward fully automated pipe fitting fabrication. However, the success of such systems depends not only on the accuracy of model reconstruction but also on the reliability of the welding process itself. Engineers must ensure that the welding parameters, including current, voltage, travel speed, and gas flow, are optimized for the specific joint geometry and material combination. The parametric reconstruction algorithm provides the geometric input, but the welding process parameters must be determined through proper welding procedure qualification (WPQ) in accordance with ASME Section IX or ISO 15614.

This study also highlights the importance of digitalization in modern manufacturing. The transition from manual welding to robotic welding requires not only advanced hardware but also sophisticated software for model reconstruction, weld seam extraction, and process planning. Engineers involved in pipe fitting fabrication should familiarize themselves with these technologies to ensure that their fabrication processes remain competitive and capable of meeting the quality requirements of modern industrial applications.