敦煌壁画的交互式线描生成技术研究

发布时间：2018-02-28 19:34

本文关键词： 敦煌壁画线描生成壁画层次结构壁画缺失内容　出处：《天津大学》2012年硕士论文　论文类型：学位论文

【摘要】：在文化遗产的保护中，敦煌壁画是一个引起大家注意的话题，而针对敦煌壁画的线描生成，又是一种特殊的保护手段。本文提出了一种交互式敦煌壁画线描生成框架，系统能在一定量的用户指导下，将一张大区域受噪声（损坏或缺失）影响的壁画，生成对应的壁画线描图，用以壁画临摹、识别和学习等多方面。本文将壁画线描生成的过程分为了三个阶段，提出了一个完整的指导用户进行壁画线描生成的框架，结合敦煌壁画的特点完成壁画线描生成的过程。首先，本文定义了壁画的结构元素及其对于壁画线描生成的重要性，采用图像分割的技术提取壁画的结构信息，本文利用全局的gPb图像分割技术抑制壁画中的全局噪声对于图像分割的影响，在用户的一定阈值控制下，得到由若干强弱不同的线条分割开的层次区域，对这些区域进行颜色标记得到完整的壁画结构信息，，作为下一个阶段的指导。其次，对于壁画线条的提取本文采用了常见的边缘检测技术，为避免壁画上噪声的影响，首先对壁画进行了平滑处理，然后使用canny算子提取了壁画的内容线条。最后，本文对于敦煌壁画中常见的壁画元素进行二进制特征描述符统计，并根据统计数据建立了壁画元素特征数据库，对于用户指定的缺失区域进入该数据库中按元素类别进行查找，推荐出一定数量的匹配壁画元素供用户进行替换。同时，本文也讨论了对现有壁画的拼接和矢量化微调的应用，提供给用户一定的修改空间。
[Abstract]:In the protection of cultural heritage, Dunhuang murals are a topic that attracts people's attention, but the line drawing of Dunhuang murals is also a special protection means. This paper proposes an interactive Dunhuang mural line drawing generation framework. Under the guidance of a certain number of users, the system can generate the corresponding line drawings of a large area mural affected by noise (damage or absence), which can be used to copy, identify and learn the murals. This paper divides the process of generating murals line drawing into three stages, and puts forward a complete frame to guide users to generate murals line drawing, and combines the characteristics of Dunhuang murals to complete the process of generating murals line drawing. First of all, this paper defines the structural elements of murals and their importance to the generation of murals line drawing, and uses the technology of image segmentation to extract the structural information of murals. In this paper, the global gPb image segmentation technique is used to suppress the influence of the global noise in the mural on image segmentation. Under the control of a certain threshold value of the user, the hierarchical region is obtained, which is divided by a number of strong or weak lines. Complete mural structure information is obtained by color marking of these areas as a guide for the next stage. Secondly, the common edge detection technique is used to extract the murals. In order to avoid the influence of the noise on the murals, the murals are smoothed first, and then the content-lines of the murals are extracted by canny operator. Finally, this paper makes the binary feature descriptor statistics of the common murals in Dunhuang murals, and establishes the mural element feature database according to the statistical data. A certain number of matching mural elements are recommended for the user to replace when the user-specified missing area is entered into the database by element category. At the same time, this paper also discusses the application of splicing and vectorization fine-tuning of the existing murals. Provides the user with certain modification space.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TP391.41;K879.41

【共引文献】