摘要
为了更加直观地表现建筑形态,提出基于最小生成树的建筑模型数字化重构方法。利用地球坐标系表示采集到的建筑信息,将射线向量引入到地球平面中,去除与平面没有关联的数据,降低数据维度;采用K-means算法确定聚类中心,以相似性为依据完成分类,提高数据的有序性;经过几何和投影变换,令重构后的模型满足人眼视觉需求;建立建筑布局图,确定节点和边集合,计算两个节点间的欧氏距离,对所有边赋予权重,权值最小的边即为最小生成树;将最小生成树的寻优过程转换为建筑模型重构过程,选用Kruskal搜索算法设置初始节点和初始边,设定迭代停止条件,确保建筑模型结构被完全搜索到,实现建筑模型数字化重构。仿真结果表明,所提方法能够提高建筑数据处理效率,重构后建筑模型与实际建筑物的相似度较高。
In order to express the architectural form more intuitively,this paper proposed a digital reconstruction method of building models based on minimum spanning tree.Firstly,the collected building information was shown by an earth coordinate system,and then the ray vector was introduced into the earth plane to remove the data that was not associated with the plane,thus reducing the data dimension.Secondly,the K-means algorithm was adopted to determine the clustering center.Based on similarity,the classification was completed,and then the order of data was improved.After geometric and projection transformation,the reconstructed model could meet the demands of human vision.Thirdly,the building layout was designed to determine the set of nodes and edges.After that,the Euclidean distance between the two nodes was calculated.Moreover,the weights were given to all edges,so that the edge with the minimum weight was regarded as the minimum spanning tree.Furthermore,the optimization process of the minimum spanning tree was transformed into the reconstruction process of the building model.Meanwhile,the Kruskal search algorithm was used to set initial nodes and edges.Finally,a terminating condition of iteration was set to ensure that the building model structure could be completely searched.Thus,the digital reconstruction of building model was achieved.Simulation results show that the proposed method can improve the efficiency of building data processing.And the similarity between the reconstructed model and the actual building is high.
作者
董扬
杨永辰
DONG Yang;YANG Yong-chen(College of Traffic Engineering,Huanghe Jiaotong University,Jiaozuo Henan 454950,China;School of Mining and Geomatics Engineering,Hebei University of Engineering,Hebei Handan 056038,China)
出处
《计算机仿真》
北大核心
2023年第3期197-201,共5页
Computer Simulation
关键词
最小生成树
建筑模型
数字化重构
聚类算法
建筑布局图
Minimum spanning tree
Building model
Digital reconstruction
K-means clustering algorithm
Buildinglayout
