A Reduction Method of Three-Dimensional Point Cloud

Recently, non-contact measurement technology has improved significantly. With the increasing of the accuracy and the speed of data acquisition of 3D laser scanners, the amount of point data has increased dramatically . 3D laser scanners generate up to thousands of points per second, which have become a burden of both computation and store of the data. It is quite important, therefore, to reduce the amount of acquire point data and convert them into formats required by reconstruction processes while maintaining the accuracy. In this paper, we presented a convenient way to solve the problem. The scattered point cloud data is first regularized and compressed by the octree structure and then reduced further according to a curvature rule. Compared with the other reduction methods, the method presented in this paper not only reduced the arithmetic complication on space and time , but also preserved the characteristic of the original object and finished the data reduction quickly. This paper presents a novel approach of point cloud reduction based on octree structure and curvature rule. The proposed method not only reduces the amount of point data and computational complexity but also makes the point cloud data be organized, which makes it easy to be traversed and searched in reconstruction process. The proposed methods are applied to different types of surfaces and the results are discussed.