基于Harris特征与NDT-ICP算法的钢箱拱预制件尺寸智检方法
Intelligent dimensional inspection method for steel box arch prefabricated components based on Harris features and NDT-ICP algorithm
王晓明 1邓璐 2史一哲 2张通 3袁通 2寇宇 4李晓 3刘宇轩2
作者信息
- 1. 长安大学旧桥检测与加固技术交通行业重点实验室,陕西西安 710064;长安大学公路学院,陕西西安 710064
- 2. 长安大学公路学院,陕西西安 710064
- 3. 陕西省交通规划设计研究院有限公司,陕西西安 710065
- 4. 铜川市交通运输局,陕西铜川 727000
- 折叠
摘要
针对桥梁建造时传统人工尺寸检测在面对海量桥梁预制件时效率低、成本高的难题,使采用地面激光扫描(TLS)技术的智能尺寸检测突破现有数据处理算法的精度与效率瓶颈,建立了基于建筑信息模型(BIM)-TLS的桥梁钢预制件尺寸智检框架,包含构件几何尺寸检测与数字预拼装2个环节;二次开发了 BIM点云化处理技术,构建了参照点云模型,采用直通滤波、统计去噪(SOR)滤波、体素化网格(VG)处理等算法预处理点云数据,实现了基于k近邻(kNN)算法的尺寸检测指标评价;通过3D-Harris特征点检测、正态分布变换(NDT)粗配准与迭代最近点(ICP)精配准提出了基于Harris特征与NDT-ICP算法的快速配准尺寸智检策略,并结合工程需求应用于某大跨拱梁组合结构钢箱拱预制件尺寸智检中.研究结果表明:采用提出的智检方法对2个相邻节段钢箱拱进行尺寸检测的最大偏差分别为1.689和1.571 mm,均满足制造偏差(小于2 mm)要求;与传统NDT-ICP算法相比,该方法将点云整体配准精度提高了 35.3％,效率提高了 61.88％,可见该方法表现高效且结果准确,促进了钢预制件几何尺寸检测智能化;基于该方法的拱肋数字预拼装监测点最大检测拼装偏差为1.953 3 mm,符合拼装偏差(小于2 mm)要求,实现了精准偏差检测,为后续桥位顺利架设提供了良好保障,且为相似结构的尺寸检测提供了参考.
Abstract
In response to the challenges of low efficiency and high cost of traditional manual dimensional inspection in face of massive bridge prefabricated components during the bridge construction,and to break through the accuracy and efficiency bottlenecks of existing data processing algorithms in the intelligent dimensional inspection using the terrestrial laser scanning(TLS)technology,an intelligent dimensional inspection framework for bridge steel prefabricated components was established based on the building information modeling(BIM)-TLS,including two links:geometric dimensional inspection and digital pre-assembly of components.The BIM point cloud processing technology was customized,and the reference point cloud model was constructed.The point cloud data were preprocessed by using the straight-through filtering,statistical outlier removal(SOR)filtering,voxel grid(VG),and other algorithms.The dimensional inspection index evaluation based on the k-nearest neighbor(kNN)algorithm was realized.Through the 3D-Harris feature point inspection,normal distributions transform(NDT)coarse registration,and iterative closet point(ICP)fine registration,a fast registration intelligent dimensional inspection strategy based on the Harris feature and NDT-ICP algorithm was proposed and applied to the intelligent dimensional inspection of steel box arch prefabricated components of a large-span arch beam composite structure in combination with the engineering requirements.Research results show that the maximum deviations of the proposed intelligent inspection method for the dimensional inspection of two steel box arches at adjacent segments are 1.689 and 1.571 mm,respectively,and meet the requirement of the manufacturing deviation(less than 2 mm).Compared with the traditional NDT-ICP algorithm,the proposed method improves the overall registration accuracy of the point cloud by 35.3％and the efficiency by 61.88％.It can be seen that the method is efficient,and the results are accurate.It promotes the intelligence of the geometric dimensional inspection of steel prefabricated components.Based on the method,the maximum inspection assembly deviation of the digital pre-assembly monitoring point for the arch rib is 1.953 3 mm,and meets the requirement of the assembly deviation(less than 2 mm).The method realizes the accurate deviation inspection.It provides a good guarantee for the smooth erection of subsequent bridge positions and a reference for dimensional inspections of similar structures.2 tabs,15 figs,30 refs.
关键词
桥梁工程/尺寸检测/三维激光扫描/NDT-ICP算法/桥梁钢预制构件/点云配准/智能化施工Key words
bridge engineering/dimensional inspection/3D laser scanning/NDT-ICP algorithm/bridge steel prefabricated component/point cloud registration/intelligent construction引用本文复制引用
基金项目
National Natural Science Foundation of China(52178014)
国家自然科学基金项目(52178104)
陕西省交通科技项目(23-59X)
中央高校基本科研业务费专项资金项目(300102212905)
出版年
2024
NETL
NSTL
万方数据