Three-dimensional laser scanning technology is a technology that can quickly and efficiently collect apparent information of objects and objects.It has been widely used in terrain surveying,cultural relics protection,mining engineering and other fields.Although the hardware development of 3D laser scanning technology has been quite mature,the theory and method of point cloud automation processing are still in the exploration stage,and many problems need to be solved.Point cloud data segmentation,feature extraction,deviation detection and reverse creation of BIM models are several key issues in the post-processing of 3D laser scanning technology.This article takes the automatic segmentation of bridge structure point cloud as the starting point.On the basis of comparative analysis of several point cloud segmentation algorithms,combined with the characteristics of bridge point cloud data structure,an automatic filtering segmentation algorithm based on point cloud projection is proposed.And deeply explored and studied the integration application of point cloud and BIM.The main research contents are as follows:(1)The system introduces the point cloud data format and its structure,analyzes and summarizes the characteristics of several point cloud formats.And the definition of point cloud segmentation and the application in different scenarios provide basic theoretical support for the subsequent algorithm research.(2)The actual point cloud is segmented for the planar model segmentation algorithm,morphological feature segmentation algorithm,and regional growth segmentation algorithm.Based on the segmentation effect analysis,the segmentation advantages and existing problems of the three algorithms are summarized.It also applies to the practical projects of point cloud on the ground flatness detection,arch rib grouting deformation detection,direct import reverse modeling and reverse modeling based on feature data.The analysis is carried out,and the problems existing in the application of point cloud at the current stage are summarized.(3)On the basis of previous research work,a filtering automatic segmentation algorithm based on point cloud projection is proposed.The whole segmentation process is mainly divided into three parts,followed by the overall point cloud segmentation of the bridge,the non-planar ground point cloud segmentation,and the bridge structural unit point cloud segmentation.In addition,contour segmentation is proposed in the feature cloud point cloud feature extraction,which further improves the degree of refinement of point cloud segmentation.The algorithm combines the idea of neighborhood query and interval density,which effectively improves the processing efficiency and accuracy of the point cloud segmentation of bridge structures.And based on the theoretical experiment,the actual project is checked,and the algorithm’s point cloud segmentation effect is verified.The bridge and its ground point cloud after segmentation will be the data foundation for the application of subsequent projects.(4)The project application research is carried out on the bridge point cloud data after the previous segmentation.It mainly includes the detection of the verticality of the point cloud of the pier column in the high altitude area,and the reverse creation of the real BIM model of the bridge.In the application of point cloud deviation detection,the overall least squares method is used to automatically fit the center line of the cylinder,which improves the accuracy of the fitting.In the reverse modeling application,the visual programming software Dynamo is used to perform feature data processing and reverse BIM model creation,and the deepening design based on the bridge reverse BIM model is realized.The automatic segmentation and application results of the point cloud show that the segmentation algorithm can not only segment the bridge point cloud of this type correctly.Moreover,the point cloud data after segmentation can be used to reversely create the real BIM model of the bridge,so as to realize the transformation from point cloud to BIM. |