Research On TSL-based Fast Inspection Algorithm For Manufacturing Quality Of Bridge Steel Beam

With the rapid development of China’s transportation industry,the scale of domestic bridge construction is constantly expanding.At the same time,bridge construction technology is also moving towards digitization,greenization,and efficiency.Currently,the size parameter detection of large bridge steel structures in China is still in the traditional manual detection stage,requiring the use of traditional measuring tools such as steel rulers,flat rulers,and plug gauges,which have problems such as large errors and low efficiency.As a new measurement technology,3D laser scanning uses 3D laser scanning equipment to scan the detection object from all directions and obtain the 3D point cloud data of the scanned object.By using effective algorithms to process the point cloud data,the required size parameters can be obtained.Therefore,replacing traditional manual measurement with 3D laser scanning technology is of significant practical significance for promoting modern bridge construction.This paper presents a method for detecting the size parameters of large bridge steel structures based on point cloud data.The specific research content is as follows:(1)Detailed introduction of the main detection indicators and existing acquisition methods for size parameter detection of large bridge steel structures,and detailed explanation of the working principle of 3D laser scanning technology,common types of point cloud data,common formats of point cloud data,commonly used 3D laser scanners,and the workflow of 3D laser scanning technology.(2)In response to the problems of high density and noise points in the raw point cloud data,this paper introduces the pre-processing method of point cloud data and some commonly used point cloud data processing software on the market.After analyzing the advantages and disadvantages of various point cloud processing software,the SOR filtering algorithm and octree downsampling algorithm provided by the open-source point cloud processing software CloudCompare are finally selected for point cloud data denoising and simplification.In addition,the plug-in developer mode of the software can be used to integrate and develop the algorithm proposed in this paper.(3)A size parameter detection algorithm for steel structures based on point cloud data is proposed.First,the region growing algorithm based on super voxels is used to segment the point cloud data of the steel structure,and the various combined steel plates of the steel structure are separated.This method has better segmentation effect than traditional region growing algorithm.Then,the point cloud data of the steel structure is gridded using the virtual grid and hash table-based point cloud partitioning algorithm.According to the index value of each grid,the position where the size parameters need to be detected can be quickly located.Finally,the plane equations of each combined steel plate of the steel structure at the detection position are calculated using the singular value decomposition algorithm.Based on the position relationship and plane equation between the planes,the corresponding size parameters can be calculated using spatial geometry principles.(4)This paper takes the steel structure of Heyong Bridge as an experimental case.First,the Trimble X7 3D laser scanner is used to collect the 3D point cloud data of the upper structure steel structure of Heyong Bridge.Then,the CloudCompare software is developed for the second time to compile the size parameter detection algorithm proposed in this paper into a function plug-in embedded in the CloudCompare software.Finally,the collected point cloud data of the steel structure is imported into the CloudCompare software developed for the second time for detection.The experimental results verify the feasibility of the algorithm proposed in this paper and achieve fast detection of some size parameters of the steel structure of the bridge.