Multi-view point cloud fusion method based on projection

Abstract

The invention discloses a multi-view point cloud fusion method based on projection, and the method comprises the steps: enabling adjacent projectors to project pure white patterns to a measured objectseparately, extracting the pixel point positions of the overlapped areas of the adjacent projectors in the image, and enabling the pixel points to be reconstructed to a three-dimensional space to obtain the point cloud of the overlapped areas; using the Kd-Tree to calculate the average shortest distance sigma of the point clouds in the non-overlapping region as a reference point interval, and then for point clouds in overlapping region, calculating nearest neighbor point of point clouds in adjacent point clouds by using Kd-Tree; combining distance and normal vector judgment, judging the overlapped points as corresponding overlapped points when the point distance is less than 2 sigma and the normal vector direction is consistent, finally calculating the distance between the corresponding overlapped points; if the distance sigma is less than the normal vector direction, performing merging by adopting progressive weighting, and determining a weighting coefficient according to the distance from the point on the three-dimensional point back projection image to the boundary of the overlapped region; and if the distance is greater than sigma, interpolating the midpoints of the two pointsby adopting natural adjacent point interpolation.

Description

technical field [0001] The invention relates to a projection-based multi-view point cloud fusion method, which belongs to the field of three-dimensional reconstruction in computer vision. Background technique [0002] The 3D measurement system based on grating projection has the advantages of simple hardware configuration, high measurement accuracy, high point density, high speed, and low cost, so it has always been a dynamic field. However, multiple cameras can effectively solve problems that cannot be solved by a single camera, and the application scenarios are more extensive. There are often common parts between the point clouds reconstructed by multiple cameras. Due to calculation errors and measurement noises, there are registration errors. The points of the common parts from two different point clouds will overlap instead of completely fitting, or even Slight delamination occurs, making subsequent application difficult. Modeling a large number of redundant points is ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the problem of partial overlapping of multi-view point clouds, the present invention proposes a projection-based multi-view point cloud fusion method

Images