三维建筑物结构化多尺度语义建模

Structural Multi-Scale Semantic Modeling of 3D Building Models

摘要: 构建三维建筑物语义模型是数字城市和智慧城市建设的需要。结构是建筑物模型区别于自然物体模型和普通人造物体模型最主要的特征，也是人类模型认知的关键。通过对三维建筑物模型的结构进行识别，可以获取原始模型的形状和构造信息，进而建立符合人类认知习惯的语义模型。本文从结构识别角度介绍了三维建筑物模型的构造语义和功能语义信息提取算法，并提出基于主体结构和细节结构的几何关联性构建建筑物多尺度语义模型的有效方法。

Abstract: 3D semantic building models conform to the construction needs of digital city and smart city. Structure is not only the main feature that distinguishes building models from those of the natural and ordinary man-made objects, but also the key factor which influences the recognition of human beings. Through structure recognition of 3D building models, we can obtain the shape and construction information of the original models, and build the semantic models that conform to the recognition habits of humans. This paper introduces the algorithms which are able to extract the constructive and functional semantics of the 3D building models from the perspective of structure recognition, and proposes the effective method to build multi-scale semantic model of 3D building models based on the geometric relationships between the main and detailed structures.