基于三角网格模型细节分析的浮雕曲面生成研究

发布时间：2018-05-28 11:07

本文选题：数字浮雕 + 频率域　；参考：《南京航空航天大学》2011年硕士论文

【摘要】：数字化浮雕自动生成是浮雕CAD研究的重要内容,其中由三维模型自动生成二维半浮雕曲面是近年来国内外浮雕CAD领域研究的热点。该方法可以以实物模型为原型,通过三维扫描仪获得模型的数字化外形,再通过一定的压缩策略,生成具有丰富细节的浮雕曲面。本文研究基于三角网格曲面细节分析的浮雕生成技术,对测量获得的三维几何模型进行频域分析或者基于光顺的细节分析,获得曲面细节的频域组成或多细节层结构。其中频域分析的方法分为基于图像频谱分析算法和直接基于三角网格频谱分析算法两种。在频谱分析中高频信息包含了更多的模型细节,基于光顺的细节分析方法中光顺前后的曲面高度差表示了模型细节。本文研究的具体内容如下: 提出了基于图像频谱分析的三维模型浮雕生成算法,该算法通过对三维模型进行重新采样,获得视线方向上的规则化深度图,然后利用图像傅立叶变换对深度图进行频谱分析,最后将模型高度的压缩转换为频谱域能量系数的压缩来得到浮雕曲面模型。在压缩过程中,通过对低频能量的压缩采用高通滤波器工具、对高频边缘压缩应用对数压缩函数等方法,使得生成的浮雕达到较好的效果。提出了基于三角网格频谱分析的浮雕生成算法。首先对三维网格模型进行频谱分析,然后通过对低频信息进行大比例压缩、对高频信息进行保真或放大处理来实现艺术多样化的浮雕。算法对高低频能量分频点的选取、浮雕高度的估算、浮雕细节调整等问题进行了详细的讨论,实例表明算法对细节丰富的模型具有较好的效果。提出了基于三角网格光顺和细节分析的浮雕生成算法。该算法通过光顺方法对模型细节进行逐次抽离,得到原始网格的基曲面与细节曲面,浮雕曲面的设计就成为基曲面的大尺度压缩和细节曲面的还原。实例表明该方法速度快,生成的浮雕效果好。文中用多个实例对提出的算法进行了算法验证,并对各算法的优缺点和应用场合进行了分析比较。
[Abstract]:The automatic generation of digital relief is an important content in the research of relief CAD, in which the automatic generation of two-dimensional semi-relief surfaces from three-dimensional models is a hot spot in the field of relief CAD at home and abroad in recent years. This method can take the physical model as the prototype, obtain the digital shape of the model through the 3D scanner, and then generate the relief surface with rich details through a certain compression strategy. In this paper, an embossed generation technique based on triangulated mesh surface detail analysis is studied. The frequency domain analysis of the 3D geometric model obtained from the measurement or the detail analysis based on fairing are carried out to obtain the frequency domain composition or multi-detail layer structure of the surface detail. The methods of frequency domain analysis can be divided into two kinds: image spectrum analysis algorithm and triangular grid spectrum analysis algorithm. In the spectrum analysis, the high-frequency information contains more details of the model, and the height difference between the surfaces before and after fairing represents the details of the model in the detail analysis method based on fairing. The contents of this paper are as follows: An algorithm for generating 3D model relief based on image spectrum analysis is proposed. By resampling the 3D model, the regularized depth map in the line of sight direction is obtained, and then the spectral analysis of the depth map is carried out by using the image Fourier transform. Finally, the compression of the height of the model is converted to the compression of the energy coefficient in the spectrum domain to obtain the relief surface model. In the process of compression, the high pass filter is used to compress the low frequency energy and the logarithmic compression function is applied to the high frequency edge compression. An algorithm of relief generation based on spectrum analysis of triangular mesh is proposed. Firstly, the spectrum of 3D mesh model is analyzed, then the high-frequency information is compressed in large proportion, and the high-frequency information is processed by fidelity or magnification to realize the diversified relief of art. The algorithm discusses in detail the selection of high and low frequency energy frequency division points, the estimation of relief height and the adjustment of relief details. The example shows that the algorithm has a good effect on the model with rich details. A relief generation algorithm based on smooth and detailed analysis of triangular mesh is proposed. Based on the method of smoothing, the basic surface and the detail surface of the original mesh are obtained, and the design of the relief surface becomes the large-scale compression of the base surface and the restoration of the detail surface. The example shows that the method is fast and the relief effect is good. In this paper, several examples are used to verify the algorithm, and the advantages and disadvantages of each algorithm and its application are analyzed and compared.
【学位授予单位】：南京航空航天大学
【学位级别】：硕士
【学位授予年份】：2011
【分类号】：J312;TP391.41

【相似文献】