Research On Optimization For Reversal Design Based On Industrial CT Sectional Images

Industrial computed tomography(CT) technology, as a main method of nondestructive testing in the reverse engineering field, which can obtain the measurement data of the objects with complex inner cavity without damaging the objects. To reconstruct the 3D imaging by using obtained measurement data, we can get 3D frame model that can directly reflect the internal structure, which facilitate subsequent product quality analysis and reverse design modification. However, due to the constraint of industrial CT scan condition and the uncertain errors brings by site operation, the mesh model may have many problems, such as the large quantities of data to stored, mesh shape is not standard and so on, all of these problems can affect the subsequent storage and finite element analysis of mesh model.To solve above problems, this paper is based on the industrial CT scan data as the research object, aiming at improving the reverse design experience. This paper studied the mesh reconstruction, mesh simplification and mesh standardization, the main research content is as follows:(1)The first key of reverse engineering steps is to achieve data conversion from two-dimensional to three-dimensional, this paper selected MC(Marching Cubes) algorithm to reconstruct the 3D mesh model required by experiment based on the analysis of the current advantages and disadvantages of 3D visualization rendering. And then taking industrial CT serial images of the hub and cylinder head as research object to verify the feasibility of the method, the experiment result shows that the MC algorithm can better reconstruction surface model of the product.(2)The model reconstructed by industrial CT serial images have huge data. To solve these problems, a simplification method named adaptive triangular mesh simplification was proposed. This method is based on the triangle collapse method, forecasting fold point position by using the area information of the fold triangle, putting the gradient increased distance threshold and angle threshold as constraint conditions to control the simplification error adaptively, and reserving the geometric features of the original triangular mesh model. The experimental result shows that, comparing with the original simplified method, this method can obtain the mesh model consistent with target mesh number, and well keep the details of the features of the original model.(3) In order to improve the triangular mesh quality of the mesh models reconstructed by industrial CT slice images, the mesh regularization algorithm combining the vertex adjustment and feature factors was presented. Introducing the improvement degree of the local mesh quality and the model feature factors as the control condition of the mesh adjustment, retaining the local detail features of the original model. Adopt the way of self-adaptive control mesh adjustment according to the normal vector threshold increasing regularity, realizing identify different curvature features automatically. Compared with the existing methods, experimental results show that the presented method can better regulate the triangular mesh while preserving the details features of the original model, and keep the geometric errors at a low level.The research achievements above provide a reference for product designers so that they can better improve the product,in addition these methods have been integrated into a prototype system of reverse engineering and the validity of the algorithms also have been verified.