Research On Automated Generation Algorithm Of BIMs Based On 3D Laser Scanning Point Cloud

In recent years,the construction industry is gradually shifting from the construction of new buildings to the maintenance of existing buildings.Combining 3D laser scanning and BIM technology,the point cloud data of the built 3D surface environment can be quickly captured and the building information can be shared.At present,The point cloud conversion BIM model is mainly implemented by importing the point cloud into the BIM modeling software manually.The process is time-consuming and information is easily loss.The automation of the point cloud generation BIM model process can solve these problems greatly,and the domestic research in this field is almost blank.Based on conventional planar buildings,this paper proposes a complete method for automatically generating BIM models from 3D point clouds.Firstly,the collected point cloud data is denoised,reduced,plane extracted,and boundary detected,and the geometric information of the plane is extracted.Then,the component type of the plane is identified,and the planes of the same component element are combined to realize the segmentation and type identification of the component elements.Finally,the parameters of the component geometry are created based on the combined plane extraction,and the BIM model is automatically generated by Revit secondary development.In this paper,the algorithm analysis and design are carried out for each step of this process.The tensor voting algorithm is used to realize point cloud denoising and the voxel method is used to streamline data.The planar point cloud is extracted based on an improved region growing algorithm that proposes the combination of spherical neighbors and K-nearest neighbors.The point cloud dimension is reduced by the planar projection method,and then the point cloud boundary contour is identified based on the ?-shapes algorithm.The component type of the recognition plane is identified by the normal direction of the design plane and the feature description of the design plane profile,and the accuracy of the plane type identification is verified by the existing type topology rules.For the surface of the conventional component type column,beam,plate,wall,door and window,etc.,the common contour edges satisfying the type property are parallel or parallel to each other,and the distance between the planes or the contour range is close to the principle.The corresponding algorithm is designed to realize the combination of the same component plane.Through the surface analysis of the model,the parameter model is used to express the component model,and the shape parameters of the conventional type components are analyzed.Based on the spatial geometry knowledge,the design algorithm extracts the corresponding component parameters from the plane combination.Based on the Revit software platform,the secondary development program picks up the component parameters,automatically creates each component model,and automatically generates the BIM model.In order to verify the feasibility of automated generation BIM method based on point cloud and related algorithm,some algorithms are implemented by Matlab programming and data experiments are carried out.The results show that the method is feasible.Finally,a brief discussion on the direction of further research in this direction is given.