Research Of 3D Reconstruction For 3D Data Set And Virtual Cutting For Model

3D reconstruction of the 3D data set and virtual cutting of the 3D model are hot research areas. The data sets got by the scientific computing and the engineering survey are often 3D data sets. 3D reconstruction of the isosurface or the 3D model of the measured object from the 3D data set can make the data analysis and processing more visual. Virtual cutting operation is an important interactive means, on the one hand, the virtual cutting operation can eliminate the mutual block of different parts on the model; on the other hand, 3D virtual cutting simulates the real operation on the object, the best operation plans can be got by the simulation. In this paper, the problems of how to achieve 3D reconstruction from three-dimensional data set and 3D model data post-processing technology are discussed. In view of the surface model containning large quantities of triangles, the mesh simplification algorithm is introduced. A virtual cutting algorithm is introduced based on the segmentation of the raw data.In this paper,there are the following innovations:(1) Aimed at the data generated by the Marching Cubes algorithm can't effectively distinguish child isosurfaces and can't effectively reflect the connective relationship among triangles, and also there has data redundancy. First, the quick sort and merger algorithm is used to merge vertexes data. According to the character of the vertices shared by the triangles, which are connecting with each other, the mutual connective relationships of the triangles are made certain. According to the character of triangles which are in the same child isosurface connect with each other, the seed algorithm is used to group the data belong to different child isosurfaces. Finally, the vertex table and triangle table are used to store the data of the same child isosurface, grouping and meshing the data of Marching Cubes algorithm are achieved.(2) Based on the theory of Marching Cubes algorithm, an improved algorithm to extract isosurface from the 3D data field is introduced. At first the 3D data set is decomposed to the topological structure of points, lines, aces and cubes. Intersection points of the lines with the isosurface are calculated at first in the 3D data field. The intersection lines of the faces with the isosurface are gained by joining the intersection points in the faces. Intersection lines are joined together to form the space polygons in the cubes. The triangle mesh of the isosurface would be got by triangling the space polygons in each cube, the triangle mesh would be registered by a vertex table and a triangle table. Based on the relationship of connection among triangles at the vertexes, the seed algorithm is used to mark the vertexes and the triangles belong to the same child isosurface. And then the individual vertex table and triangle table register vertexes and triangles belonging to the same child isosurface.(3) The control points are selected from the initial mesh triangles. The first control points are the center points of the initial mesh triangles, each first control point is corresponding to one initial mesh triangle. The second control points are the vertexes of the characteristic lines. Each control point has a under control triangle in the mesh. Each mesh triangle has a first control point set and a second control point set. Two algorithms is introduced. For the first algorithm, the distance between the control point to the under control triangle is the error between the initial mesh and the simplified mesh. The triangle collapse operation should meets that the distance between the control point to the under control triangle is less than the threshold value. For the second algorithm, the energy function is defined for the mesh. The energy function is used to control the simplification process. The hybrid mesh simplification process, which is the combination of the triangle collapse and point delete, is introduced. Each simplification operation should meet the decline of the mesh energy value.(4) A new algorithm for virtual cutting on the 3D model is introduced. First, by the property that the data belong to the same child object are 6 connective in the 3D data set, the data belong to different child objects are marked, so the data of the child object are grouped. The cube set is generated for the data belong to the same child object. The virtual cutting face is separated to a group of line segments. Two algorithms are introduced to calculate the intersection of the line segment and the cube set. For the first algorithm, the octree is generated for the cube set, by calculating the intersections among the octree of the cube set and the line segments, the intersection points of the virtual cutting face and the cube set are got. For the second algorithm, the line traversal algorithm calculates the intersection. The cutting layer data is formed in the data set. By using the improved seed algorithm, the data that have been cut are regrouped, the surface model is rebuilt by using the MC algorithm, and then the algorithm for virtual cutting on the 3D model is achieved.