基于偏微分方程的目标体表面光滑研究

发布时间：2018-08-18 14:37

【摘要】：目标体表面的曲面重构技术在医学,工业,地质等领域有着广泛的应用,属于信息可视化研究的重要领域。其中基于空间散点的曲面重构方法可以总结为三个主要步骤:构建网格,提取特征和曲面光滑处理。得益于多年来许多研究者在相关领域的不懈钻研,许多高效而严谨的算法和思想被提出并投入到实践应用当中。不过这些方法依然有许多不足的地方。目前针对地质数据没有一种可以在保持曲面趋势不变的情况下构建出光滑曲面的算法。在地质数据处理领域也没有一种效果特别好的特征提取算法。本文对目标体表面的曲面重构进行了研究,并针对高程模型曲面的网格构建,特征提取和曲面光滑处理进行了深入探索,实现了一套曲面重构的完整流程。本文的主要创新包括:1.总结了传统高程模型曲面特征提取的不足之处。本文针对这些不足,引入基于曲率的特征提取方法,并结合流水法特征提取算法提出了一种融合特征的算法。这种算法可以减少特征碎片化,提取出更完整的特征信息。2.针对传统曲面光滑方法会改变曲面趋势的缺点,本文将特征约束加入基于偏微分方程的曲面光滑处理。结果不但可以保留原曲面特征,还能使曲面光滑度达到所需水平。综上所述,本文针对曲面重构领域中的高程模型曲面重构方法进行了改进和实现。改善了曲面网格插值效果,使曲面的插值结果更贴近原始曲面的趋势。改善了特征提取的算法,使特征提取的结果更全面。将基于偏微分方程的曲面光滑方法与特征提取相结合,改善了曲面光滑的结果。还实现了一个曲面重构可视化系统。
[Abstract]:The surface reconstruction technology of object surface has been widely used in medical, industrial, geological and other fields, which belongs to the important field of information visualization. The method of surface reconstruction based on spatial divergence can be summarized as three main steps: grid construction, feature extraction and smooth surface processing. Many efficient and rigorous algorithms and ideas have been put forward and put into practice because of the unremitting research of many researchers in related fields over the years. However, there are still many shortcomings in these methods. At present, there is no algorithm to construct smooth surface while keeping the trend of surface unchanged for geological data. There is also no feature extraction algorithm in the field of geological data processing. In this paper, the surface reconstruction of the object surface is studied, and the mesh construction, feature extraction and smooth surface processing of the elevation model surface are deeply explored, and a complete process of surface reconstruction is realized. The main innovations of this paper include: 1. The shortcomings of surface feature extraction of traditional elevation model are summarized. In this paper, a feature extraction method based on curvature is introduced, and a feature fusion algorithm is proposed in combination with the pipelined feature extraction algorithm. This algorithm can reduce feature fragmentation and extract more complete feature information. In view of the disadvantage that the traditional surface smoothing method can change the surface trend, the feature constraints are added to the surface smoothing process based on partial differential equation in this paper. Results not only the original surface features can be retained, but also the smoothness of the surface can reach the required level. To sum up, the surface reconstruction method of elevation model in the field of surface reconstruction is improved and implemented in this paper. The effect of surface mesh interpolation is improved, and the interpolation result is closer to the trend of original surface. The algorithm of feature extraction is improved and the result of feature extraction is more comprehensive. The method of surface smoothing based on partial differential equation is combined with feature extraction to improve the result of surface smoothing. A visualization system for surface reconstruction is also implemented.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;O175.2

【参考文献】