【作者】 刘雁；

【导师】 耿兆丰；

【作者基本信息】 东华大学 ， 控制理论与控制工程， 2003， 博士

【摘要】 迄今为止，计算机辅助服装设计(CAGD)主要在绘图和计算层面上实现了对服装设计师的支持，提高了设计效率。但对于更深层次的服装设计智能思维活动的模拟则缺乏实质性的进展。随着科学技术的发展和市场需求变化的日益加快，传统服装CAD技术在服装生产应用中存在的不足日渐彰显。一方面，现今的服装款式设计软件都是针对平面款式效果图设计的，不能直接体现款式的三维效果，需要通过样板设计和样衣制作才能得到款式的三维效果；另一方面，目前服装CAD系统是以模拟传统服装设计过程而开发的，其交互式的操作过程相当烦琐，对使用人员的专业水平要求高，学习和熟练过程也很长，极大地限制了系统的推广和使用。 对于服装设计而言，最理想的方法就是将款式设计要求与三维款式效果直接联系，实现款式设计要求驱动下的三维服装效果自动生成，但现在国内外在这方面的研究尚属空白。因此，本文从三维和智能两方面对服装款式设计展开研究，提出了基于原型的三维服装款式智能设计方法，为实现三维服装款式智能设计系统奠定了理论基础。 本文针对三维服装款式智能设计，主要展开了如下研究工作： 1、对服装设计思维与模型展开研究。首先探讨了服装设计思维过程和思维机理，针对服装设计过程中常用的思维方式，建立了相应的思维模型。其次，在分析了服装设计从构思到三维款式生成的全过程的基础上，提出了以原型法作为研究三维款式智能设计的理论基础。最后对该方法在整个服装生产过程中的功能和作用进行了探讨。 2、对服装设计知识进行了提取与表达。分别从服装设计和知识属性两个角度对服装设计知识进行了研究，并根据服装设计知识的特点选择了恰当的表示方法，充分表示了各种类型的服装设计知识，为今后设计问题的求解奠定了完备的知识基础。 3、针对三维服装模型的智能控制开展研究。在分析了各种三维建模技术的基东华大学博士学位论文础上，提出了有利于对模型形态变化进行可预见性控制的参数建模方法，并就如何运用该方法对服装各组成部件进行曲面拟合问题进行了探讨。另外，本文运用Matlab建立了三维人体模型及服装模型，提高了建模效率。 4、对服装款式设计知识与三维原型参数的关系展开研究。主要研究了三维服装模型参数对模型形态的影响关系、模型参数彼此之间的制约关系以及三维模型参数与服装设计知识的关系，并将这些关系运用产生式规则和算法进行了表达。以衣领和裙装为实例，实现了服装设计知识对三维款式形态的控制。 5、鉴于三维服装款式智能设计是对服装设计知识、三维模型控制知识以及彼此之间的关系知识融合的过程，本文构建了三维服装款式智能设计系统结构，分析了智能设计的过程和关键技术问题，研究了如何对各种知识进行有效地组织、管理和运用，实现多知识融合的三维服装款式的智能生成，并对三维服装款式智能设计进行了仿真。 本文研究工作的创新点主要体现在以下几个方面: 1、鉴于国内外针对服装设计思维的相关研究很少，从而使得服装智能设计系统由于缺乏理论基础而难以实现的现状，本文针对服装设计思维过程和机理开展了深入研究。首次提出了基于约束满足的服装设计思维模型、基于借鉴的服装设计思维模型、基于类比综合的服装设计思维模型和基于原型的服装设计思维模型。并在基于原型的服装设计思维模型基础上，提出了基于原型的三维服装款式智能设计方法，为将计算机辅助服装设计从对款式图形的辅助绘制引向深层次的对服装设计思维的智能模拟奠定了基础。 2、服装设计知识往往以经验性知识的形式存在于设计师的头脑中，缺乏系统的提取和表示。为了实现计算机辅助服装款式智能设计，本文首次运用了多种方法对服装设计知识进行表示:运用框架法对服装设计知识的整体结构和层次进行表示，建立了服装设计知识模型;运用产生式规则对变化丰富、经验性强的款式设计知识进行了表示;针对服装款式设计中的模糊性知识，特别提出了通过建立模糊设计知识与三维服装款式形态要素的关系列表的方法对其加以表示。 3、国内外针对三维服装辅助设计的研究一般只注重实现三维款式图形的交互式控制，而没有涉及对三维服装款式图形的智能控制，本文针对这一空白开展研究，摘要运用参数三次样条曲线和双三次曲面建立了三维服装模型，为实现三维服装款式的智能变化奠定了基础。 4、本文首次对服装款式设计知识与三维模型控制知识之间的关系进行研究，建立了一种新的关系知识模型，并将这种知识表达为产生式规则和求解算法，从而使款式设计知识延伸到对三维模型参数的控制，为实现款式设计知识控制的三维服装模型智能变化奠定了基础。 5、本文提出的三维服装款式智能设计是一个多知识融合的过程。本文提出以服装部件为对象单位，运用面向对象技术，建立了服装款式设计知识库，对各种知识进行了有效的组织与管理。并提出了基于框架表示的款式设计知识的模糊匹配算法，以及按照匹配度进行冲突消解的策略。同时指出了三维服装款式智能设计是对多目标节点的求解过程。通过研究问更多还原

【Abstract】 Conventional computer aided garment design (CAGD) system improves the pattern design and style design efficiency greatly with the excellent function of calculating and drawing, but it is incapable of simulating the creative design thinking of fashion design. With the development of technology, the disadvantages of the conventional CAGD system become noticeable. The first disadvantage of the conventional CAGD system is that it is developed for designing two-dimensional garment style, and the designer can’t see the three-dimensional style effect of their design unless the pattern are made and the sample dress are sewed. The second disadvantage of the conventional CAGD system is that it is aimed at copying the conventional garment design process and realizing interactive operation (which means the user has to tell the computer what to do step by step), so the system has to be used by the professionals and takes them a long time to master it. These disadvantages holdback the application and popularization of the system.The most effective way for quick response of fashion design is to enter the requirements of design and see the effect of 3D style on computer screen directly. This is what we called intelligent design process of 3D garment, and it is a new direction the conventional CAGD may head for. However there are few researches done in this field. In this article, we dedicate to researching on the intelligent design of 3D garment for the first time, and bring forward the prototype-based method of 3D garment intelligent design.The main content of this thesis are as follows:1. We researched on the fashion design thinking and model. Analyzed the process and mechanism of fashion design thinking, and set up the model of several common used methods of thinking in fashion design. Based on the analysis of the whole process through the conceiving of fashion style to the producing of 3D garment effect, we brought forward the prototype-based method of 3D garment intelligent design, and introduced the procedure as well as the function of the new method during the whole process of garment manufacture.2. We extracted and expressed the knowledge of fashion design. First, we analyzed the attribute of the knowledge of fashion design from the point of view of the process of fashion design and the property of knowledge. Based on this, we chose the fit expressing method of knowledge to express the different style of fashion knowledge. Thus thefashion knowledge can be expressed well, and a maturity knowledge foundation for solving the fashion design problem can be established.3. In order to control the deformation of the prototype of 3D garment automatically, we brought forward a method to set up the 3D garment prototype by using the technique of cubic spline and bicubic surface patch, and discussed detailedly about how to apply the method to set up the 3D prototype of different parts of the garment. In addition, we chose the Matlab as the modeling tools in our research work and improved the modeling efficiency greatly.4. We researched the relationships between the knowledge of fashion design and the value of the parameters of 3D prototype. First, we studied the relations between the shape of the 3D model and the parameters of the 3D model, as well as the relations between the parameters themselves. Then we expressed these relationships into algorithms and generated rules. At last, we chose collar and skirt as the examples to demonstrate the influence of the changing of values of the parameters on the style of the 3D garment.5. The process of 3D garment intelligent design is the process of fusing all the knowledge gained above. First, we built up the structure of the system of 3D garment intelligent design, and analyzed the key technology and procedure of the intelligent design. Then we studied how to organize, manage and utilize these knowledge. At last we simulated the intelligent design of 3D garment.The innovative works of this thesis are as follows:1. There are few researches done on the thinking activity更多还原