【作者】 刘哲铭；

【导师】 毛昕；

【作者基本信息】 东北大学 ， 机械设计及理论， 2010， 硕士

【摘要】 作为先进制造技术的一个重要组成部分,反求工程已从最初的原型复制技术逐步发展成为支持产品创新设计和新产品开发的重要技术手段。与这一趋势相对应,反求工程重建的CAD模型已不再是传统的实体模型,而是更具创新设计能力、支持产品全生命周期的特征模型。在面向特征造型的反求工程技术背景下,本文以产品设计意图与测量意图理解为着眼点,以重建支持产品创新的参数化特征模型为目标,围绕反求工程CAD建模中的特征识别、约束识别、全局优化及参数化等关键技术展开研究。点云截面曲线的成功参数化建模对于提高反求工程模型重建的精度和效率,实现逆向创新设计都有很好的应用价值；同时,对于丰富计算机图形学研究具有理论价值。本文所做的工作主要有以下几个方面：(1)系统研究了常见曲线特征的数学表达和性质以及曲线特征在几何约束条件下的拟合技术。给出了截面点云特征识别的基本含义,并且提出了一种新的特征识别准则,能够实现一段点云数据的基本特征的识别。(2)本文给出了常见曲线之间各种约束的表达形式,并在此基础之上建立了一种约束识别准则,可以实现各平面几何特征之间的约束识别。(3)在进行截面形状特征约束重构时,可以将由点云数据拟合得到的各段曲线,在满足各段之间的约束条件下进行整体约束逼近,从而构造整个截面形状特征。本文通过建立一通用的全局优化模型,来解决满足各段之间的约束下进行整体约束逼近的问题。并且尝试了利用MATLAB提供的最优化计算工具箱中的fmincon函数进行最优化计算,可以得到较理想的结果。并且从逼近误差和约束误差分析了全局优化的结果。(4)本文研究了基于曲线的参数化设计方法,给出了Bezier曲线的参数化表达形式,并对自由曲线的等距偏移进行了一定的研究。(5)最后在总结全文工作的基础上,指出了有待进一步研究的内容和方向。更多还原

【Abstract】 As an important component of advanced manufacturing technology, reverse engineering gradually developed into an important technical means of product innov-ation and new product development from the initial prototype support replication technology. Corresponding to this trend, Reverse Engineering CAD model recon-struction of the entity is no longer a traditional model, but more innovative design capabilities, supporting for product lifecycle feature model. In the context of reverse engineering technology for feature modeling, this paper start the study to focus the understanding of the product design intent and the measurement intent, to rebuild the parametric feature model supporting for product innovation, to focus the key technologies in reverse engineering CAD modeling, such as feature recognition, constraint recognition, global optimization, parametric modeling and so on.Unorganized point cloud curve parametric modeling is very valuable both for improving the accuracy and efficiency of reconstruction and innovative design and for enriching computer graphics research. This major work done mainly in the following areas.Firstly, both the nature and the mathematical expression of the common curve characteristics and the fitting techniques of Curves by the numbers was researched. The basic meaning of point cloud feature recognition and a new feature recognition criteria was introduced. Basic characteristics of point cloud data can be identified successfully by this method. Rebulided parameters was keeped consistent with the original design parameters as much as possible.Secondly, in this paper, the expression of various constraints between common curveis was introduced. On this basis, a constraint recognition criteria was established in this paper, which can identify the constraints between the various geometry characteristics.Thirdly, rebuilding constraints of the cross section, each curve which point cloud data can be fitted to can have an overall constraint approximation to construct the shape of cross section in the conditions of meeting the constraint between paragraphs. A general the global optimization model was given to solve the problems of an overall constraint approximation in this article. The fmincon function of MATLAB Optimization Toolbox was used of solving optimization calculation. That can get effective results. The results of global optimization from approximation error and constraint error was analysed.Fourthly, in this paper, parameter design method based on the curve was researched. The parameters expression of Bezier curves and the approach to satisfy the constraint relations were given. Equidistant offset of the free curve has been researched in this paper.Finally, on the basis of summarizing the work, the content and direction of the further research were given.更多还原