Rock mass structural plane intelligent interpretation method based on point cloud and GPU technologies

Abstract

The invention discloses a rock mass structural plane intelligent interpretation method based on point cloud and GPU technologies. Collected rock mass laser point cloud data are processed as follows: step one, normal vectors of all points in the point cloud data being calculated based on GPU parallel computing and a linear regression method; 2, identifying a structural plane feature point set by using an improved region growing method; 3, extracting a structural plane boundary from the final structural plane feature point set by utilizing an improved convex hull scanning method; and 4, calculating the trace length, occurrence and spacing of the automatic interpretation structural plane based on the contour points. According to the method, the top-speed operation of massive point cloud datacan be realized, the efficiency bottleneck of identifying the structural plane based on the point cloud in the past is broken through, the rock mass structural plane information is accurately, quicklyand efficiently acquired, and meanwhile, the boundary of the structural plane can be described with high precision, so that the accurate trace length of the structural plane is calculated.

Description

technical field [0001] The invention belongs to the technical field of rock mass structural plane extraction, in particular to an intelligent interpretation method for rock mass structural planes based on point cloud and GPU technology. Background technique [0002] Due to the developed structural planes inside the rock mass, the rock mass shows a certain structure, and also directly affects the mechanical properties of the rock mass. How to efficiently and accurately obtain the geometric information of rock mass discontinuities has become a hot and difficult problem in rock mass engineering. In actual engineering, due to the advantages of convenient measurement means and accurate measurement results, the manual on-site contact measurement method is widely used in structural surface measurement. However, this method has some disadvantages, such as low contact measurement efficiency, potential safety hazards and incomplete measurement results. [0003] Three-dimensional las...

AI Technical Summary

Problems solved by technology

[0004] There are roughly three types of existing point cloud-based rock mass structural plane measurement methods: the first type of method needs to manually select more than three points, and extract structural plane information based on this, which is highly subjective and inefficient

The second type of method is to use the TIN network model. This type of method has the following disadvantages: 1) part of the point cloud will be lost after triangular meshing, and there is some distortion in describing the surface details; 2) when there are many discrete points in the point cloud data, or When the point cloud data is in a closed state, the process of triangular meshing will produce more errors, and even many deformed triangles will appear

Although this type of method avoids the shortcomings of the above two types of methods, it takes about ten minutes to process a million-level point cloud; and in actual engineering applications, rock mass point cloud data is often tens of millions of levels. At this time, the traditional The structural surface recognition algorithm has been far from meeting the efficiency requirements, which has hindered the practical application of structural surface extraction based on point clouds

Images