Three-dimensional entity model reconstruction method based on dense point cloud data

Abstract

The invention provides a three-dimensional entity model reconstruction method based on dense point cloud data. The dense point cloud data is acquired by a laser 3D scanner; 3D laser scanning can rapidly acquire three-dimensional coordinate data of the surface of a measured object in a large-area and high-resolution manner; a rapid and high-precision modeling method is matched, so that the characteristics of rapidness, non-contact, initiative, accuracy and the like are highlighted, data obtained in real time has the characteristics of high accuracy, high density and the like, reconstruction ofa three-dimensional model becomes more convenient, a digital model of an actual object can be created in the virtual world, and rapid and high-accuracy three-dimensional entity model reconstruction based on dense point cloud data is achieved; according to the semantic-based feature extraction method, a large amount of working time is saved, the semi-automatic mode is adopted for operation, the feature extraction precision and accuracy are effectively improved, the cost is greatly reduced, and the precision of the reconstructed three-dimensional solid model is obviously improved.

Description

technical field [0001] The invention relates to a three-dimensional solid model reconstruction method, in particular to a three-dimensional solid model reconstruction method based on dense point cloud data, and belongs to the technical field of point cloud three-dimensional reconstruction. Background technique [0002] 3D laser scanning, also known as real scene reproduction, is a technological revolution in the field of surveying and mapping after the GPS technology. 3D laser scanning technology breaks through the traditional single-point measurement method, and has the unique advantages of high efficiency and high precision. Through high-speed laser scanning measurement method, large-scale high-resolution rapid acquisition of target outer surface information, in order to establish a three-dimensional solid model of the object Excellent technical means are provided. With the wide application of computer technology in various fields, reverse engineering has become one of th...

AI Technical Summary

Problems solved by technology

But the defect of RANSAC's method: the optimization consideration is single, and the constraints between the geometries are not taken into account, so the splicing between the reconstructed model geometries is not harmonious, and gaps are prone to occur.

[0012] First, it is difficult to quickly obtain the three-dimensional coordinate data of the surface of the measured object with large area and high resolution in the existing technology, and there is also a lack of fast and high-precision modeling methods. The technology does not have the characteristics of fast, non-contact, active, and accurate. Without the characteristics of high precision and high density, there are many difficulties in the reconstruction of 3D models, which is not conducive to creating digital models of actual objects in the virtual world;

[0013] The 2nd, prior art is in the acquisition of point cloud data, and the acquisition cost of point cloud data is high, and the acquisition speed of point cloud data is slow, can't obtain the dense point cloud data that satisfies the follow-up processing of the present invention to go, can't be high-precision three-dimensional The solid model reconstruction lays the foundation; point cloud data registration accuracy is low, and the degree of automation is low; a large number of human-computer interaction drawing extraction methods in the existing technology consume a lot of working time and high work cost;

However, in the actual use of the Lawson method, when more data is encountered, the efficiency of the network structure will slow down. When the range of the point cloud is not convex or there is an inner ring, illegal triangles will be generated. The Bowyer-Watson method randomly points The time complexity of insertion and triangulation is large, and the efficiency of finite element mesh generation is very low;

[0015] Fourth, the defect of various methods using RANSAC in the prior art is that when the sampling data information of a specific shape is less, that is, when the proportion of the entire point cloud data is small, the search in the entire point cloud is time-consuming and inefficient; The RANSAC method of the prior art has a single optimization consideration and does not take into account the constraints between the geometries, so the connection between the reconstructed model geometries is not harmonious, and gaps are prone to occur; the main disadvantage of the prior art reconstruction based on small-area shapes is that Large-area missing data cannot be completed, because the reconstruction effect of this method is strongly dependent on the small-area standard shape template. The geometry template used is for the description of the small-area point cloud shape, and the small-area shape matching process takes a long time. Inefficient method

Images