Method for detecting multi-dimension characteristic line of noise grid model

Abstract

The invention relates to a method for detecting multi-scale characteristic lines, which aims at a noise mesh model. The method comprises the following steps: constructing a discrete multi-scale representation of a three-dimensional mesh model; obtaining respective optimal local scales of each vertex on the three-dimensional mesh model according to the probability estimate; calculating the curvature information of each vertex on the three-dimensional mesh model under the corresponding optimal local dimension; detecting characteristic points according to the curvature information, and using thecharacteristic points to generate the characteristic lines. The method can effectively overcome the effect of noise in the data, and rapidly acquire the detection results.

Description

Multi-scale Feature Line Detection Method for Noisy Mesh Model technical field The invention relates to the fields of computer vision and graphics, in particular to a method for detecting characteristic lines of a three-dimensional grid model. Background technique With the continuous development of the 3D scanning system, it is easy to obtain a large amount of data and a high-precision 3D model. In order to obtain a more reliable digital geometry processing method based on 3D model data, the detection of feature lines is an important part of the process of 3D model processing. It has become one of the research hotspots. Generally speaking, the characteristic line refers to a curve that has nothing to do with the viewpoint and reflects the geometric change of the surface of the 3D model. It is defined as a line composed of vertices whose main curvature on the 3D model has a positive maximum value or a negative minimum value along its corresponding main direction. Includin...

AI Technical Summary

Problems solved by technology

For the 3D mesh model obtained from the original scanning data, due to the inevitable noise in the data, if the traditional method of directly detecting the feature line is used to detect the feature line, the detection result will contain many redundancy due to noise. Curves, it is difficult to distinguish them from the required characteristic lines by simple post-processing operations

Images