A Filtering Method for Inhomogeneous Scattered Point Cloud Data

Abstract

The invention discloses a filter method of nonuniform unorganized-point cloud data. The method comprises the steps of index organizing the unorganized-point cloud data of a gob, and organizing the point cloud data to a data model of a cubic regular grid; determining a grid approximation plane S by selecting the gravity core of three adjacent grids with most grid measured points; solving a shortest distance h from each measured point in the grid to the grid approximation plane S by utilizing a normal vector n; determining a filter range, and eliminating a noise point; adjusting the size of the regular grid for the filtered point cloud data, and repeating the steps to iterate the filter. By adopting the filter method, the loud noise and partial low noise can be effectively eliminated, and the elimination error rate of the nonuniform data can be reduced.

Description

technical field [0001] The invention relates to the technical field of point cloud data processing, in particular to a filtering method for uneven and scattered point cloud data. Background technique [0002] At present, 3D scanning technology has improved the accuracy of deformation detection of 3D measurement objects and provided an effective basis for formulating and optimizing later support schemes. Due to the complexity of the measurement environment and the influence of the scanner's own errors, the acquired original point cloud data contains a lot of noise, resulting in a large error between the 3D reconstruction model and the geometry of the actual measurement object. In order to obtain a more accurate 3D reconstruction model, in the analysis Based on the original point cloud data and noise sources, it is necessary to effectively denoise the point cloud data. [0003] The noise of goaf point cloud data mainly has three forms: regional floating noise, isolated noise ...

AI Technical Summary

Problems solved by technology

When dealing with uneven and scattered point clouds, the smoothing denoising method may mistake the non-noise area with sparse sampling points as noise, resulting in enlarged model holes and weakened geometric features after smoothing; while the method of removing bad points has an obvious effect in removing large-scale noise , but cannot remove small-scale noise and weaken the geometric features of the model. The above two types of denoising methods are difficult to adapt to noise of different scales and maintain the original geometry of the goaf when dealing with uneven and scattered point cloud noise.

Images