张宏鑫, 方雨桐, 刘木水, 董斌智. 复杂建筑平面图纸的空间布局融合对偶识别方法[J]. 武汉大学学报 ( 信息科学版), 2021, 46(9): 1354-1361. DOI: 10.13203/j.whugis20210323
引用本文: 张宏鑫, 方雨桐, 刘木水, 董斌智. 复杂建筑平面图纸的空间布局融合对偶识别方法[J]. 武汉大学学报 ( 信息科学版), 2021, 46(9): 1354-1361. DOI: 10.13203/j.whugis20210323
ZHANG Hongxin, FANG Yutong, LIU Mushui, DONG Binzhi. Dual Recognition Method of Spatial Layout Fusion for Complex Architectural Plan Drawings[J]. Geomatics and Information Science of Wuhan University, 2021, 46(9): 1354-1361. DOI: 10.13203/j.whugis20210323
Citation: ZHANG Hongxin, FANG Yutong, LIU Mushui, DONG Binzhi. Dual Recognition Method of Spatial Layout Fusion for Complex Architectural Plan Drawings[J]. Geomatics and Information Science of Wuhan University, 2021, 46(9): 1354-1361. DOI: 10.13203/j.whugis20210323

复杂建筑平面图纸的空间布局融合对偶识别方法

Dual Recognition Method of Spatial Layout Fusion for Complex Architectural Plan Drawings

  • 摘要: 矢量化的复杂建筑平面图纸受各类图元的干扰,其空间布局识别难度非常高,然而在5G基站建设、智能家居以及AR/VR(augmented reality/virtual reality)中广为应用。提出了一种结合栅格图像和矢量表示的高性能融合对偶识别方法。首先,采用建筑平面图纸的栅格化表示,提取建筑图纸的主要发展方向,构建分割的若干个封闭区域空间。然后,采用建筑平面图纸的矢量化表示,利用空间识别的有向性和邻接特性,创新地提出使用半墙数据结构进行墙体几何位置和拓扑关系的计算,从而根据空间对偶原理高精度地整体重建墙体布局。最后,获取包含各类墙体布局的矢量建筑平面图数据集,并进行方法的实验验证。实验结果证明了所提出的先空间后墙体的融合对偶识别算法对于各类建筑模型类型均具有可用性和有效性。与传统先墙体后空间的识别方法相比,所提方法具有更高的鲁棒性,受特定建筑图纸类型的干扰更少。

     

    Abstract:
      Objectives  The spatial layout of vectorized complex architectural drawings is widely used in 5G base station construction, smart homes, and AR/VR(augmented reality/virtual reality).
      Methods  To solve the difficulty of identifying the spatial layout of drawings caused by various irregular drawing elements, a high?performance fusion recognition method combining raster image and vector representation is proposed. First, the rasterized representation of architectural plan drawings is used to extract the main development direction, and several enclosed spaces are constructed. Then, the vectorized representation of architectural plan drawings is used and according to the directional and adjacency characteristics of space recognition, the half?wall structure is innovatively proposed to calculate geometric position and topological relationship. The wall layout can be rebuilt as a whole with high precision according to the principle of space duality. Finally, the proposed method and the traditional methods are analyzed on the vector building plan data set of various wall layouts.
      Results  The experimental results show that the proposed fusion dual recognition algorithm is usable and effective for various types of building models.
      Conclusions  It has higher robustness and is less interfered by specific architectural drawing types.

     

/

返回文章
返回