The Surface Matching And The Error Analysis Based On The Triangular Mesh

Reverse engineering is an important design and development method inindustrial manufacturing. Reverse engineering could analyze, study and improvesystemically on the existing physical product and sample, and then develop moreadvanced products and part models, which are revised and re-designed. The digitalproduct designing and manufacturing can achieve in variety place based on thereverse engineering. Right now, reverse engineering is widely applied to many fieldsin industry production and social life such as creative product design method,mechanical manufacture, restoration archeological findings, drawing designing ofproducts and medical science.Data acquisition, data process and surface matching are key technology ofreverse engineering. Digital method of product can realize data collecting of productsurface, and acquire the geometry information of the part by the particularmeasurement equipments and methods. These lay the foundation for complex surfacemodeling, evaluating, improving and manufacture. Data processing technology canobtain the correct and complete measured data by converting the data format, smoothfilter, data compaction and segmentation. These can get ready for the follow-upmodeling work. Surface matching technology is a data process that unites the data indifferent coordinate systems to the same coordinate system. It aligns data in diffrentspatial positions together, and stitches and matches data in different eyeshots. Theseprovide a correct way for workpiece measurement, precision manufacturing andspringback control. Surface matching technology is the core technology for datamerging and surface quality control, and it is also the basis of error evaluating.In this paper, the key technology of data acquisition, data process and surfacematching are studied based on the triangular mesh.Then error analysis method ofsurface after matching is discussed in detail. With the help of OpenGL and VisualC++, this paper has made the surface registration algorithm and error cloud displaycome true. The main research contents and conclusions in the paper are as follows:1.With the help of3-D scanner based on the structure light technology,3-Dcoordinate information of workpieces are obtained by means of nondestructivemeasurement. Because of the influence of system error，measure environment and thesurface quality of workpiece, the obtained data will contain certain noise, whichwould inevitably disturb the following data processing. For this reason, weightedmedian filter method is employed to filter point cloud data of the forming surface. In addition, in view of the properties of structural flexibility and simple topology of thetriangle mesh, we describe the target surface in the form of triangular mesh to discretethe surface model. We read data and display it in the shape of point, line and planerespectively, through the VisualC++basic programming language and the powerfulgraphing functions of OpenGL. Then orientate point cloud data according to geometrytransform principle.2.The z-axis method proposed in the text is employed to coarsely aligned theforming surface and the target surface, which could narrow the differences betweenthe positions in the great degree. Computational method of α in both theory andexperiment are proposed. These verify the effectiveness of this method.3.The new method of closest point searching is proposed based on triangularmesh. The features of absolute value function and cubic parabola function areanalyzed，and objective function according to absolute value function is constructed.On the basis of z-axis method, two points cloud data are matched, the registrationtests of surface model with different shapes verify the effectiveness of the algorithm.The z-axis method aligns the data to general location, and taking it as the initial valuecan improve the primal ICP algorithm both in speed and accuracy and achieve highaccuracy registration.4.Based on the discretized triangular mesh, analyze the error of the data aftermatching by the projection interpolation method. By the measuring point projectingon the triangular mesh of target surface and determining point of intersection, theerror value is obtained by interpolation calculation in triangle polygon meshes. Thewhole error and local error in a certain section after matching are shown. These canrealize the whole error of surface matching, and can check the error of point cloud inconcrete position after surface matching. Based on the matching algorithm and errordisplay method, data processing results of the software developed by myself andgeomagic qualify commercial software have been compared. Results show that myown software approximates the latter in accuracy and the algorithm can be applied inthe practice to deal with the surface matching problem.In brief, the surface matching based on triangular mesh can provide a newmethod to workpiece positioning, quality testing and springback compensation. Thecalculation and analyses in theory provides theoretical basis for question of theproduce physically. The study has important theoretical value and practicalsignificance.