The Realization Of A High Precision Reconstruction Algorithms Of B-spline Curve Based On Cloud Data

With the development of 3D scanning technology, 3D space data is extracted by the laser scanning technology,which the structure and feature of the object is shown, and the original model is restored by these data points, it is the key problem in present study.Firstly, the application and situation of inverse engineering and reconstruction of curve are introduced. The basic theory of B-spline curve is expounded which include definition, generating principle and classification, and nature of B-spline curve is analyzed in the paper.Secondly, data model is researched based on 3D space data points, and Multi-level data structure is established, and a more accurate boundary point judgment rule is put forward.A new fast extraction algorithm of eight part is presented based on k-nearest neighbors, find out the nearest 30 points from the current point, and the algorithm of boundary points quick searching is designed in any two continuous area of 2D plane, and the algorithms of quick searching is achieved. The shortcoming of four parts that repeated calculation of each data point is improved, and the problem of data points missing in the area of large curvature is overcome in the algorithm, and the accuracy of extraction and the speed of calculation are improved; and the sorting algorithm of boundary points is put forward, and the orderly arrangement of boundary points according to the nearest distance is realized.Then, on the basis of orderly boundary points, an algorithms of B-spline curve based on control points inverse, in which the problem of node parameter setting is solved, in order to be convenient for calculation, set the parameters of data points according to standard, and parameter is calculated, closed treatment of curve is realized, control points are calculated which has the same coordinate with the original data points, and analyze the special control points, chase algorithm is used to calculate.Finally in VC++2008 compiling environment, the uniform three times B-spline curve reconstruction of the boundary contour data model of the bunny rabbit is achieved by programming, and gives the running result of the algorithm, the precision between original boundary points and fitted curve is analyzed. The experimental result is compared with result in MATLAB, proves that the design of the algorithm is feasible and the reconstruction accuracy is higher. The problems of sharp corner, deformation and smoothness that result from sparse data points and large curvature in fitting are overcomed in the algorithm. The algorithm is a good method and reconstruction precision is higher, and it has a certain reference value for curve and surface reconstruction.