Font Size: a A A

Improvement And Procedure Realization Of Data Compression Algorithm In Reverse Design

Posted on:2008-10-26Degree:MasterType:Thesis
Country:ChinaCandidate:Z Z ShiFull Text:PDF
GTID:2178360212997494Subject:Mechanical design and theory
Abstract/Summary:PDF Full Text Request
1. IntroductionThe traditional product realization usually is from the conceptual design to the pattern, remakes the product, we call it the forward project, but the reverse-engineering generates the pattern according to the components, then to product. It is a kind of technology taking the advanced product equipment, the type, the software or the phantom as the object of study, applying modern design methodology, the production engineering, the material and the related specialized knowledge to do the systematic analysis and research, the exploration grasps its key technologies, then develops the similar more advanced product , is a union of analysis method and the application technology which is absorption of advanced technology . At present, majority research and application related the reverse engineering all concentrate in the geometry shape, namely the reconstruction of product material object CAD model and the final product manufacture aspect, which is called the reverse-engineering of material object. in order to adapt the advanced technique of manufacture development, it needs by certain way to transformer a type into the CAD model, then using the computer assistance manufacture, the fast prototype manufacture and the fast mold, the product data management and the computer integration manufacture system and so on to process or manage.at the same time, along with the modern test technology development, gaining the material object geometry information fast and precisely comes true. At present, this kind of related technology which gains product mathematical model from the type and obtains the new product, has become a research and application hot spot in the CAD/CAM domain, and develops into a relatively independent discipline domain. Under this significance, the reverse engineering may be defined : The reverse engineering is the digitized technology which transformer tmaterial object into CAD model,the geometry model reconstruction technology and the product technique of manufacture,this article refers to this concept. The reverse-engineering as a kind of project practice technology, widely applies in the machine design. along with the optical technology development, non-contact scanning equipment has gradually substituted for contact survey - - three coordinate instrument (CMM). the non-contact survey may in a short time gain massive primitive curve surface data message , may enhance the precision of curve reconstruction in the reverse-engineering .But simultaneously it also has the shortcoming which overcomes with difficulty: Because the machine vision survey is the spacing survey, so enhance the fitting precision of boundary and the range curvature change widely, only reduces the gap between the scan line, increases the data point quantity, which can cause the data point quantity to increase suddenly, cause the curve and surface fitting computation load, How therefore to delete the redundant data point, guarantee curved surface fitting precision which already becomes the bottleneck in the reverse-engineering .This article provide an effective compression method, based on eight forked tree's inhomogeneous three dimensional method of lattice and the procedure realization.2. Octree and its application in the data compressionOctree is from Quadree which promotes a kind of 3D grid construction of data forms to the three-dimensional space. Its basic thought is to divide the 3D space region by the three dimensional grid, each small cube has one or many attribute data. Octree 's shape of tree structure to decompose in the space on has the very strong superiority. Eight forks the tree model to become a cube size three-dimensional space division eight ancient divination symbols to limit, each ancient divination symbol limits has the same time and the spatial order of complexity in, when some ancient divination symbol limits the object attribute same or may regard as same no longer subdivides, otherwise on limits this ancient divination symbol the again segmentation is eight ancient divination symbols limits. Through the circulation recursion division means, cause each individual all to belong to the identical attribute or to achieve the stipulation up to the tolerance (eight forks tree's resolution) .This article proposed the algorithm model foundation is eight fork tree structure, eight forks the tree space division model is surrounds the space with a cube sequence which a curved surface (data cloud) occupies, and regards it as eight forks the tree model the root node, divides the cube the size same eight sub-cubes, each sub-cube is regarded as the root node the sub node, from this the modeling space recursion segmentation is 2 power of exponent stature cube, this method is the grid even segmentation law, and already was widespread applied in the computer graphics and the imagery processing. It may regard as is a dendrite structure, according to the practical application difference, the segmentation standard is dissimilar, in carries on in the entity boundary fitting, these completely must delete in entity interior or exterior grid unit, but those contain boundary point the grid unit only then to be able to carry on the segmentation.This article proposed the algorithm sets up the principle take eight forks as a foundation, through the data point triangle division, calculates the data point the spot law arrow, the spot principal curvature, estimated approximately the curved surface curvature change, according to the grid inner point law arrow, the spot principal curvature standard deviation carries on the net the uneven segmentation.3. Inhomogeneous three dimensional grid data compression algorithm based on eight forks trees and its procedure realizationIn this article the data processing method is stores up the data point in the certain size three dimensional grid, then carries on the grid the segmentation, regarding all data points falls in the identical grid, using the average, the weighted average or is takes the intermediate point the method to select a characteristic data point to represent in the entire grid the spot.The algorithm concrete mentality calculates each data point place spot law vector and principal curvature, determined a cloud the outside envelope cuboids, according to the curved surface shape and the compression ratio determined the initial grid size and determined eight fork tree initial grid unit, joins the initial grid to the grid tabulation, takes a grid from the grid tabulation to look every time in grid data point quantity whether is bigger than 1, if is bigger then calculates in the grid the data point law vector sum principal curvature standard deviation whether has a standard deviation to be bigger than the given value, then acts according to the inhomogeneous grid segmentation law to carry on the grid to subdivide, afterwards erases the spatial subunit grid, finally selects with the grid in the mean point law arrow included angle ,The small spot achievement must retain the data point, erases in this grid other data points.According to in the question object-oriented analysis model foundation which establishes in the analysis stage designs in the establishment question solution space the question space class; Then through the analysis user's concrete operation requirements, designs the establishment user interface class in the question space class foundation, the task management class and the data management class; Finally, carries on the specify to a kind of design, and carries on the reexamination according to the related standard compilation design instruction booklet, completes to the question object-oriented design modeling. This article has completed a kind of foundation in this foundation, namely has sealed three dimensional data point all attributes and the operation Point kind, has sealed all cuboids attribute and in view of the cuboids operation Cuboids kind, and Application kind; Is responsible for the user interface the plan, the data circulation as well as controls the news transmission respectively the CProjectDlg kind, provides an input dialog box to the user ,the CInputBox kind, to the user demonstrated the dialog box CAboutDialog kind, is responsible to work the thread foundation, time the system resources assignment and the initialization as well as the procedure withdrawal clean-up CProjectApp kind. In Visual under the C++ formidable programming environment support, through a kind of mutual transfer, has realized the inhomogeneous three dimensional grid data compression algorithm programming, and has given the friendly user interface, and will insert large-scale reverse-engineering software like UG, CATIA for the future and so on to build the good foundation.This article will study the design the algorithm integration, greatly will enhance the compression the data precision, will enhance the working efficiency greatly. Therefore the reasonable promotion will have the very good application prospect, enhance the domestic mechanical profession, the university and the research unit's design development speed precision plays the vital role. May take the characteristic point in particular after the processing data point, next step carried on the parameterization for us based on the ACIS platform the prototype automatic production and the probabilistic model revision reconstruction has provided the parameter. From the data point obtains gain the characteristic point which to the data point compression, until in carries on the prototype automatically the reconstruction and the revision, a data file delivering craft department row of craft processes the product, this is a complete product performance history, this article completes is in this project the most essential part, therefore can design the very important function in the mechanical profession reversion.
Keywords/Search Tags:reverse design, data compression, Octree, 3D grids, non-uniform
PDF Full Text Request
Related items