B-spline surface fitting method and device based on dense point cloud

Abstract

The invention provides a B-spline surface fitting method and device based on dense point cloud, and the method comprises the steps: collecting the three-dimensional point cloud data of a scanning partthrough a laser scanner; carrying out filtering and denoising processing on all the collected point clouds; analyzing curve fitting times suitable for the point cloud data by utilizing a learning curve; performing feature point extraction on all input point clouds to obtain sparse feature points; carrying out B-spline curve fitting under a UV coordinate system by using the sparse feature points,and carrying out error comparison with a line segment of original point cloud data linear interpolation fitting; continuing to select a new feature point from the curve segment with the large error, and carrying out re-fitting and error analysis; repeating until the curve error reaches the precision to form a curved surface grid; splicing the reconstructed curved surface patches to form an integral scanning part model; and carrying out modeling and texture mapping on the scanned part model. Dense point cloud can be changed into sparse points through a feature point selection algorithm, and a high-precision curved surface can be fitted with the least data points.

Description

technical field [0001] The invention relates to a B-spline surface fitting method and device based on dense point clouds. Background technique [0002] When using 3D point cloud data for design and processing, or using 3D point cloud data for display, it is first necessary to establish a 3D model with a relatively accurate surface smoothness for the target object. At present, the commonly used method is to use machine vision and laser sampling to collect pictures of laser reflection from different angles of the target. The point cloud in the picture reflects the surface information of the target, and these pictures are matched and stitched to form a complete 3D model. When collecting data from different angles, multiple cameras can be set at different angles, and single or multiple cameras can also be used to collect data from different angles. But no matter which of these two methods, it involves the speed and accuracy of the reconstruction model. The reconstruction speed...

AI Technical Summary

Problems solved by technology

However, in practical applications, it is found that unless there is an accurate angle measurement device, the user is not sensitive to the angle, and it is difficult to determine the angle accurately; the size of the target object is difficult to accurately determine, especially in some applications where the target object needs to be replaced frequently. A lot of extra work is required, and professional equipment is required to accurately measure irregular targets

Measurement errors lead to camera position setting errors, which will affect the speed and effect of acquisition and synthesis; the accuracy and speed need to be further improved

Images