Three-dimensional point cloud file loading method based on automatic segmentation and spatial indexing

Abstract

The invention discloses a three-dimensional point cloud file loading method based on automatic segmentation and spatial indexing. The method comprises the following steps: acquiring a three-dimensional point cloud file to be processed; automatic point cloud segmentation is carried out on the three-dimensional point cloud file, and the automatic point cloud segmentation is to read and segment the three-dimensional point cloud file line by line according to a pre-designed point cloud segmentation mode; and performing spatial indexing and dynamic loading on the three-dimensional point cloud file. Compared with a traditional three-dimensional point cloud file segmentation loading mode, the method has the advantages that the automatic segmentation of the ultra-large three-dimensional point cloud is realized through the automatic segmentation of the three-dimensional point cloud file; according to the three-dimensional point cloud file processing method and device, the three-dimensional point cloud is divided into three-dimensional point clouds, efficient spatial indexing of the three-dimensional point clouds is achieved through spatial index library construction and dynamic loading, splicing and standardized format output of the partitioned point clouds are achieved, the processing efficiency of the three-dimensional point cloud file is improved, hardware expenditure needed by processing is reduced, and the method and device are particularly suitable for ultra-large type three-dimensional point cloud files.

Description

technical field [0001] The invention relates to the technical field of three-dimensional point clouds, in particular to a method for loading three-dimensional point cloud files based on automatic segmentation and spatial indexing. Background technique [0002] 3D point cloud is a disordered point set formed by computer-aided calculation by recording a large number of dense point coordinates, reflectivity, texture, panorama and other information on the surface of the measured object using the principle of laser rangefinder. Because of its high precision, it can efficiently, quickly and accurately obtain the surface data of real-world collection objects, and it has deep applications in smart cities, smart construction sites, land acquisition and resettlement, land planning, geological exploration, engineering quantity estimation and other directions. [0003] Different from the 3D laser point cloud in the industrial field, the point cloud file in the engineering field is huge ...

AI Technical Summary

Problems solved by technology

[0004] The first way is to directly load the entire file, such as the display method and device, equipment and storage medium of the point cloud image (authorization announcement number: CN109472852B) is exactly to load directly, but this loading method has higher requirements for computer hardware, and often Failed to load due to insufficient memory

Moreover, for the 3D point cloud data loaded by this method, in the subsequent point cloud data processing, the amount of irrelevant calculations required for positioning, cutting, etc. is huge, which wastes computer resources and the calculation efficiency is also very low.

[0005] The second method is to manually divide the point cloud file into blocks, and then load the corresponding point cloud files into blocks according to the actual needs of the business. Although this method solves the problem that the complete 3D point cloud cannot be loaded due to insufficient memory , but it is still impossible to quickly extract the corresponding point cloud area according to the specific coordinates required by the actual business

The former has high requirements for computer memory and is not suitable for loading super large 3D point cloud files; the second method requires the design of more cumbersome file naming rules, which is not easy to manage the 3D point cloud index library, and the indexing efficiency is low

Images