Abstract: Compared with concrete structures, steel structures offer advantages such as high strength, short construction periods, and low-carbon environmental benefits, making them increasingly widely used in infrastructure projects such as large bridges. To ensure structural performance, it is essential to accurately assess the manufacturing quality of steel components prior to installation. However, at present, such detection still heavily relies on manual operation, with problems such as large site occupation and low detection efficiency. To address this, 3D laser scanning technology was employed, and the detection and evaluation study of prefabricated component manufacturing quality was carried out based on high-precision 3D point cloud models. Firstly, aiming at the complex environment of prefabricated beam yards, optimization principles for 3D laser scanning measurement station positioning and on-site implementation methods were proposed. Subsequently, preprocessing of the collected point cloud data was carried out, including registration, denoising and down-sampling, so as to improve data quality and processing efficiency. Finally, automatic inspection and evaluation methods for spatial geometric dimensions and surface flatness were proposed based on algorithms including Alpha Shape, random sample consensus (RANSAC), and principal component analysis (PCA). The proposed method has been successfully applied on the standard steel box girder in the Shenzhen-Zhongshan Bridge project. The results demonstrate that 3D laser scanning and 3D point cloud model can provide strong support for the manufacturing quality assessment of prefabricated components. The proposed detection method has advantages such as good automation, high efficiency, and visualized results, which can offer effective assistance to inspectors to carry out manufacturing quality control and management for prefabricated components.

Key words: bridge engineering; manufacturing quality assessment; prefabricated component; 3D laser scanning; point cloud data; steel box girder

摘要： 钢结构相比混凝土结构具有强度高、施工周期短、低碳环保等优势，因而在大型桥梁等基础设施项目中得到日益广泛的应用。为确保结构性能，需在安装前对钢构件的制造质量进行精准评估。然而，现阶段此类检测仍高度依赖人工操作，存在场地占用大、检测效率低等问题。采用三维激光扫描技术，基于高精度三维点云模型开展预制构件制造质量的检测与评估研究。首先，针对预制梁场的复杂环境，提出了三维激光扫描测站选位的优化原则及现场实施方法；随后，对采集的点云数据进行配准、去噪和下采样等预处理操作，以提升数据质量和处理效率；最后，基于Alpha Shape算法、随机抽样一致性算法和主成分分析算法，提出了空间几何尺寸和表面平整度的自动检测与评估方法。该方法在“深中通道”工程中标准钢箱梁上成功应用，结果表明: 三维激光扫描和三维点云模型可以为预制构件制造质量的评估提供有力支撑，所提出的检测方法具有自动化程度好、效率高、结果可视化等优势，能够有效辅助检测人员开展预制构件的制造质量控制与管理工作。

关键词: 桥梁工程；制造质量评估；预制构件；三维激光扫描；点云数据；钢箱梁