基于混凝土组合箱梁残缺三维激光点云的自动逆向建模方法

An Automatic Inverse Modeling Method Based on Incomplete 3D Laser Point Clouds of Concrete Composite Box Girder

针对虚拟预拼装过程中,施工现场条件与构件本身性质对混凝土桥梁构件三维激光扫描及其后续逆向建模质量产生不良影响的问题,以混凝土组合箱梁为背景,提出残缺三维激光点云的自动逆向建模方法。该方法利用实测点云与设计数据最佳匹配的数学优化问题提高自动性,并采用k-d tree和最小生成树(MST)等数据结构以及算法提高计算效率和鲁棒性,建立了以特征点三维坐标集合为最终逆向建模结果的通用方法。采用该方法对单根长约30 m的混凝土组合箱梁进行逆向建模并提取构件尺寸,同时与传统测量方法实测值进行比较,结果显示该方法尺寸检测误差均在1%以内,测量精度可以满足现场实时处理数据和获取结果的需求。

During virtual pre-assembly,the field condition and the characteristics of concrete components of a bridge can exert adverse effect on the 3D laser scanning results and the subsequent inverse modeling quality.In this paper,an automatic inverse modeling method based on incomplete 3D laser point clouds of concrete composite box girders is proposed.The method harnesses the optimization of matching as-designed data to as-built data to improve automobility,and uses K-dimensional(K-d)Tree and Minimum Spanning Tree(MST)to enhance the efficiency and robustness,and develops a universal method based on aggregate of characteristic points′coordinates.The method was utilized to inversely model a 30 m-long concrete composite box girder and to extract the sizes of its components.Compared with the sizes obtained by using conventional measurement method,the proposed method can generate more accurate measurements,with errors less than 1%,which meet the needs of processing real-time data and getting immediate results.