| Three-dimensional reconstruction which is based on the reverse engineering, is a hot and difficult question of the artificial intelligence, machine vision and virtual reality and other cutting-edge areas and is also one of the major challenges human in basic research and applied research. Three-dimensional reconstruction is an important research branch of the image processing technology, simulating directly the way human eyes processing scenery, which has a very broad application prospect in 3D terrain simulation, military reconnaissance, 3D communication, e-commerce, etc.An important issue of the three-dimensional reconstruction is feature points reconstruction, which first acquires objects feature points,and then achieves object surface reconstruction through some tectonic methods.This paper mainly introduces the triangulation of space scattered point cloud in three-dimensional reconstruction techniques,and study how to generate high-quality triangular mesh triangulation and reduce the complexity of compute time and space.This paper studys the factors which influence the quality, space and time complexity of generating triangular mesh through analyzing the existing algorithms,and proposes a method of processing space messy the point cloud directly which combines divided and incremental and used the octree model in the triangulation of point cloud segmentation. finally experiments show that this triangular mesh quality and time and space complexity through this paper algorithm processing is better than the traditional algorithm,and the triangular mesh result is very uniform intuitive and clear , truely reproduce the three-dimensional model of the object. |
