Application Of Point Cloud Feature Regularization Based On Field Force Algorithm In Deformation Monitoring

With the rapid development of China’s economic construction,various large and super-large buildings with complex structures have emerged as the times require.At the same time,the subway industry has transferred the way of transportation from the ground to the underground,which has improved the traffic jams of cities.During use,buildings and subway tunnels will appear deformation,such as tilting and displacement and so on,which will cause safety accidents and endanger people’s lives and property.Therefore,it is very important to conduct periodic safety monitoring on them.Considering that the traditional methods such as total station and leveling instrument have been difficult to meet the measuring requirements of deformation monitoring,3D laser scanning technology as an emerging space mapping technology has many advantages such as high efficiency,high precision and non-contact measurement,and its application in the field of deformation monitoring can greatly improve the monitoring efficiency.Therefore,based on the laser point cloud data and aiming at the shortcomings of the existing algorithms,a fast and effective point cloud feature extraction algorithm is proposed and a feature regularization model is established in this thesis.The practicability of the proposed algorithm is verified by two engineering examples,and the conclusion of deformation monitoring is finally obtained.The main efforts done in the thesis are:(1)Analyzing the working principle and the source of scanning error of the 3D laser scanner;introducing the pre-processing process and implementation method of point cloud data,which provides a data basis for subsequent feature extraction.(2)Aiming at the shortcomings of the existing algorithms,a feature extraction algorithm with improved field force criteria is proposed.The local k-d tree is established to search the k neighbors by constructing the spatial dynamic grid,then the least square method is used to fit the micro-tangent plane and project,and the field force criteria is improved according to the characteristics of folding edges,so as to extract the feature points.Experimental results show that compared with the existing algorithms,the proposed algorithm can extract feature points quickly and accurately with fewer broken lines,fewer redundant points and intact boundary characteristics.(3)On the basis of feature extraction,point cloud feature regularization model is proposed.A formula is deduced by the vector deflection angle and Euclidean distance to select the joint points,and the ordering steps are given according to the complexity of point cloud structure,then the improved cubic B-spline algorithm is used to fit the feature points to obtain boundary lines to achieve the feature regularization of the point cloud.Through qualitative and quantitative analysis of the Erqi bridge of the Yangtze river,it is found that the boundary lines after feature regularization are smooth and continuous,and with higher accuracy,the average deviation of fitting is 0.0674 mm.(4)Based on the feature extraction algorithm and the regularization model,the deformation monitoring analysis of two engineering examples with different structures is carried out.By matching the corresponding points of the two phases,the overall deformation of the northwest corner of the building is analyzed,according to the angle change of the folding edge,the inclination of the building is calculated and the inclination trend is analyzed.For the subway tunnel,the axis is used to extract 20 tunnel sections,and the section circle is fitted by the least square method,then the overall deformation of each section is analyzed,at last the local deformation of the section is calculated by intercepting corresponding points.