Research on Point Cloud Simplification Method for Steel Components Based on Mobile Laser Scanning

In response to the challenges of high outlier noise, high redundancy, and inaccurate feature information extraction in the initial point cloud of large structures such as bridges, this article proposes a point cloud simplification method for steel components based on mobile laser scanning. Firstly, statistical filtering techniques are used to remove outlier noise points, and voxel down-sampling prepocessing is performed on the point cloud data; Divide the point cloud data into uniform spatial grids, calculate curvature descriptors on a per small grid basis and optimize the grids; By calculating the average angle between curvature and K-neighborhood normal vectors, the key points are divided into feature points and described by using the Fast Point Feature Histogram (FPFH) algorithm. Through the analysis of experimental results on a large assembled component of a certain bridge, it can be concluded that under high simplification, this method, which has better performance compared to traditional simplification algorithms, can effectively avoid the occurrence of hole phenomena and has achieved good performance in simplification speed and accuracy. Therefore, this method has broad application prospects in pre-assembling steel components.