基于子图拓扑扩展的徽派建筑快速建模

发布时间：2018-04-27 23:02

本文选题：徽派建筑 + 自动建模　；参考：《合肥工业大学》2014年硕士论文

【摘要】：大规模场景建模是当前计算机虚拟现实领域研究的热门课题,并且是涉及到诸多学科的复杂问题。通过虚拟仿真建模技术实现对建筑场景的还原,对传统古建筑的宣传、保护和传承具有重要意义。然而在目前存在的建模理念中,尚未有可应用的建模技术涉及大规模的古建筑场景自动化建模。本文在对徽派建筑分析和研究的基础上,对当前各种模块化、参数化建模方法进行了总结分析,提出了一种基于子图扩展生成徽派建筑拓扑结构图的建模方法。首先,以用户简单的拓扑输入为初始信息,通过关联规则对拓扑子图进行扩展,自动地生成建筑场景的拓扑结构图,同时根据各类约束规则推理出各个节点的模块属性,然后对整个场景的拓扑图遍历,生成场景描述文件。最后,渲染得到徽派建筑的场景模型。本算法显著改善了传统大规模场景自动建模方法的建模建筑结构单一的问题,并且有效提高了建模精度和速度。关联规则是用来描述数据集中项目之间的关联特性,也是子图扩展算法的基础。本文通过改变频繁项集的链接方式和压缩事物项集的方式改进了Apriori挖掘算法,并实验对比了改进前后的算法效果。最后,对本文采用的基于哈希表和三元组的拓扑结构图的存储方式进行了分析,并与传统存储方法进行了对比,说明了该存储方式的优越性。
[Abstract]:Large-scale scene modeling is a hot topic in the field of computer virtual reality, and it is a complex problem involving many disciplines. Virtual simulation modeling technology is used to restore the architectural scene, which is of great significance to the propaganda, protection and inheritance of the traditional ancient buildings. However, in the existing modeling concept, there is no applicable modeling technology involving large-scale modeling of ancient architecture scene automation. Based on the analysis and research of Huizhou architecture, this paper summarizes and analyzes various modularized and parameterized modeling methods, and puts forward a modeling method based on the extension of sub-graph to generate Huizhou architectural topology chart. First of all, with the user's simple topology input as the initial information, the topology subgraph is extended by association rules, and the topology structure diagram of the building scene is automatically generated. At the same time, the module attributes of each node are deduced according to various constraint rules. Then the topology of the whole scene is traversed to generate the scene description file. Finally, render the scene model of Huizhou architecture. This algorithm significantly improves the single problem of building structure modeling based on the traditional automatic modeling method of large-scale scene, and effectively improves the precision and speed of modeling. Association rules are used to describe the association characteristics between items in a dataset, and are also the basis of an extended subgraph algorithm. In this paper, the Apriori mining algorithm is improved by changing the linking mode of frequent itemsets and compressing the itemsets of things, and the effect of the improved algorithm is compared experimentally. Finally, the storage method of topology structure diagram based on hash table and triple is analyzed, and compared with the traditional storage method, the superiority of this storage method is explained.
【学位授予单位】：合肥工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TU-87;TP391.9

【相似文献】