Method for shape-preserving and streamlining sampling data of physical surface

Abstract

The present invention provides a shape-preserving simplification method for object surface sampled data, belonging to the technical field of reverse product engineering. The shape-preserving simplification method is characterized by: taking sampling points and a near neighbor point set thereof as surface partial samples, and obtaining normal information of object surface sampled data based on a principal component analysis method; calculating a normal included angle mean value set phi between the sampling points and a topological near neighbor point set thereof; taking the normal included angle mean value set phi as a clustering object, estimating a natural clustering number of the normal included angle mean value set phi based on a gap statistic algorithm; performing clustering for the normal included angle mean value set phi through a k-mean adaptive clustering algorithm to obtain a natural clustering result of the object surface sampled data; and setting simplification threshold values for data sampling points of each cluster according to different shape features of the sampling points so as to achieve adaptive simplification of the sampling points. The shape-preserving simplification method of the present invention can quickly perform adaptive clustering of the object surface sampled data to realize intra-cluster simplification, so as to achieve shape-preserving simplification of the whole point cloud data.

Description

technical field [0001] The invention provides a shape-preserving and streamlining method for surface sampling data of a physical object, which belongs to the technical field of product reverse engineering. Background technique [0002] Surface reconstruction is the core technology of product reverse engineering. In order to improve the accuracy of surface reconstruction, the sampling density of the physical surface is forced to increase continuously, resulting in a large number of sample points contained in the scattered point cloud. Therefore, in the process of surface reconstruction, in order to improve the reconstruction efficiency, It is necessary to perform conformal reduction on the huge point cloud data. [0003] At present, the commonly used point cloud reduction algorithms mainly include triangular mesh division method and point cloud reduction method based on differential geometric quantity estimation. In the academic journal "Journal of Xi'an Jiaotong University"...

AI Technical Summary

Problems solved by technology

This method can effectively retain the surface characteristics of the original model, but the operation is complex, and the triangular meshing of the point cloud data reduces the algorithm efficiency and cannot be applied to large-scale point cloud data.

In the "Efficient simplification of point-sampled surfaces" published in the collection of academic papers "In Proc.IEEE Visualization 2002" 2002, 163-170, PaulyM proposed to apply clustering method, iteration method and particle simulation method to point cloud data The simplification of the clustering method is relatively fast, but the average error is relatively large; the average error of the iterative method is small, but it is not convenient to control the density of sampling points and the preprocessing process is more complicated. The particle simulation method overcomes the above two The shortcomings of the algorithm, but the algorithm does not run efficiently

[0004] To sum up, the current data reduction method for physical surface sampling has the problems of inaccurate reduction results and low reduction efficiency. Therefore, it has become an urgent need for those skilled in the art to provide a point cloud data reduction method with high reduction accuracy and efficiency. Technical issues resolved

Images