Three-dimensional reconstructing method of three-dimensional laser helical scanning point cloud

Abstract

The invention discloses a three-dimensional reconstructing method of a three-dimensional laser helical scanning point cloud. The method comprises the following steps: establishing a point cloud data structure containing a 3D coordinate, related angle information and rotary time of a point; storing the scanned helical line point cloud data according to the established point cloud data structure, and reordering the helical line point cloud data according to the related angle information and the rotary time of each point; sectioning the reordered helical line point cloud data according to the rotary time to acquire the single-contour line point cloud with the same number with the rotary time; optimizing each single-contour line point cloud, and then connecting according to a principle whether the related angels of the upper single-contour line point cloud and the lower single-contour line point cloud are the same so as to form a 3D entity model. By use of the method disclosed by the invention, all point cloud data participate the computation, the 3D morphology of the point cloud is reserved to the greatest extent, and the 3D reconstructed entity is matched with the point cloud 3D morphology, is free from holes; an accurate 3D reconstructed entity is formed so as to provide basis for the subsequent data utilization.

Description

technical field [0001] The invention relates to the technical field of point cloud data processing, in particular to a three-dimensional reconstruction method of a three-dimensional laser helical scanning point cloud. Background technique [0002] There are many sources of point cloud data, such as 3D laser scanners, photogrammetry, and three-coordinate measuring machines. Point cloud data is a scattered topological structure, a kind of scattered point cloud, also known as an unorganized data set, and there is no internal connection between points. The purpose of 3D reconstruction of point cloud data is to find a certain mathematical description form, construct a triangular mesh model with vertices and interconnected topological relationships, and use this as a basis to analyze, optimize, modify and draw the mesh itself. How to reconstruct the scattered point cloud into a 3D model is a popular research direction, but the number of point clouds obtained by scanning the measu...

AI Technical Summary

Problems solved by technology

Amenta et al. proposed the Crust algorithm based on the three-dimensional Voronoi diagram in 1998, and then proposed the Cocone algorithm to further optimize the Crust algorithm, but this type of algorithm is very computationally intensive and inefficient; Adamy et al. Sub-UmbrellaFilter method

Such algorithms all involve the calculation of Delaunay triangulation or Voronoi diagram of point cloud, which has a relatively large amount of calculation and low efficiency.

[0005] However, the above scheme still has the following defects: 1) The time and space complexity of the algorithm, that is, the efficiency problem. For the huge amount of point cloud data obtained by laser scanning, the three-dimensional reconstruction by triangulation method does not consider To scan the logical relationship between point clouds, it takes a lot of time to traverse all point clouds and triangulate them, which is often unacceptable in practical applications

2) Using triangulation for 3D reconstruction, it is necessary to calculate the convex hull of the point cloud, and then gradually approach the external 3D shape of the point cloud. The final 3D reconstruction solid model depends on the 3D distribution of the point cloud. If the point cloud has depressions or If the point cloud is sparse and uneven, the 3D reconstructed solid model will not match the reality

[0008] However, the above scheme still has the following defects: using the 3D reconstruction method based on radial basis function, the 3D reconstructed solid model will have holes, especially when the collected 3D point cloud is sparse and uneven, then the 3D reconstructed solid model will be It is in a state of being riddled with holes, which is too different from the real object surface, and the sharp features cannot be recovered

Images