【作者】 贾成；

【导师】 王启富；

【作者基本信息】 华中科技大学 ， 数字化设计与制造， 2007， 硕士

【摘要】 在产品协同开发中,复杂三维CAD模型给网络传输带来沉重负担,导致协同产品开发效率降低。因此,提高三维CAD模型的压缩效果,减轻网络传输负担成为当前学者研究的热点。本文结合国家自然科学基金“面向协同产品开发的CAD模型压缩方法和累进传输技术”(编号:50375058),从CAD模型中的自由曲线曲面入手,对产品模型的压缩算法展开研究,具体工作如下:曲线曲面是CAD产品模型的主要组成部分之一。为了取得较好的压缩效果,针对自由曲线曲面控制点压缩问题,采用一种基于差分脉冲编码调制(Differential Pulse Code Modulation, DPCM)原理的曲线曲面编码器,在一定程度上降低了自由曲线曲面控制点信息的冗余。为了实现DPCM编码器的优化,在保证预测效率的前提下,结合自由曲线曲面控制点的分布特点,借鉴最佳线性原理,通过对线性预测系数进行优化,采用一种最佳线性预测策略,为后续熵编码奠定了基础。为了实现CAD模型的精确表达,且能实现较好的压缩效果,提出一种自由曲线曲面无损压缩算法。首先,针对自由曲线曲面控制点坐标值是IEEE标准下单精度浮点数的特点,引入浮点数尾数差值编码方法,并结合最佳线性预测策略,实现尾数差值的预测,然后通过熵编码进一步提高压缩比,从而实现自由曲线曲面的无损压缩。为了获得CAD模型的高压缩比,提出一种自由曲线曲面有损压缩算法。首先依据最佳线性预测策略获得预测误差,然后按用户给定精度对预测误差进行量化,同时将量化值反馈进行下一步的预测,最后对量化值进行熵编码,从而实现自由曲线曲面的有损压缩。在上述理论研究成果的基础上,结合三维CAD模型进行了实践,并开发了针对产品模型中的自由曲线曲面编码压缩组件,将其运用到华中科技大学自主研发的轻量化组件InteVue中,获得了较好的压缩效果。更多还原

【Abstract】 3D CAD complex models lead to low efficiency for data exchange and data transmission in a network-based collaborative development environment. Compression of the 3D CAD models is a hot research topic nowadays. Supported by NSFC (No:50375058) named" the CAD model compression method oriented to collaborative development and progressive transmission technology", this dissertation analyses the character of free-form curve and surface, and proposes geometry compression algorithm about 3D CAD models .The contributions are followed:Curve and surface is main part of model. To get better compression, a Differential Pulse Code Modulation (DPCM) coder about free-form curve and surface is proposed. The coder with prediction and quantizer,followed with entropy coding,reduces the redundant of the control points of the free-form curve and surface.Optimizing the prediction about the control points is key technology in DPCM coder. Considering efficiency of prediction and character of the control points of the free-form curve and surface, optimal linear prediction achieved by optimizing the coefficient of the prediction, is realized by theory of Minimum Mean Square Error.To get the accurate data of CAD models and achieve better compression ratio, a new algorithm of lossless compression about free-form curve and surface is proposed . Mantissa coding with optimal linear prediction is introduced according to the character of floating-point values in IEEE standard, followed by entropy coding, as a result, the control points’coordinates in free-form curve and surface are compressed in completely lossless manner.To get high compression ratio , a loss compression algorithm is proposed .The scheme uses DPCM coder with optimal linear prediction for control points , coupled with a scalar quantizer which precision appointed by user, followed by bit-plane arithmetic entropy coder.Based on theories mentioned above, those approaches are applied in some 3D models .Furthermore , compression module of free-form curve and surface in the 3D models is developed to apply to the lightweight system named InteVue developed by Huazhong University of Science & Technology and achieves good compression ratio.更多还原