摘要
针对三维建筑群模型简化难题,该文提出一种基于视觉认知理论的模型群组聚类概括方法。该方法利用道路要素对场景进行粗划分;然后利用方向、面积、高度等空间认知要素及其拓扑关系约束进行精分类,使其符合城市形态学特征;采用Delaunay三角网和边界追踪综合算法进行模型合并概括,并对模型进行分层存储。应用典型城市建筑群模型进行验证,结果表明:该算法简化效率高,分类结果符合人的认知习惯,并且通过聚类概括过程中各阈值的自适应控制,对于不同的模型,概括层次能够达到相对统一。
Aiming at the problem of simplification of 3D building group model,this paper proposes a generalized clustering method based on visual cognition theory.The method uses road elements to make rough classification of the scene,and then uses spatial cognitive elements such as direction,area and height to classify the scene and make it conform to the morphological characteristics of the city.Delaunay triangulation network and boundary tracking algorithm are used to model merging,and the models are stored in different layers.The algorithm is verified by the application of typical urban building group model.The experimental results show that the algorithm proposed in this paper has high efficiency,the classification results conform to people′s cognitive habits;by the self-adaption of all thresholds in the process of clustering generalization,the generalization levels of different models can be relatively unified.
作者
李朝奎
方军
李慧婷
吴柏燕
陈爱民
LI Chao-kui;FANG Jun;LI Hui-ting;WU Bai-yan;CHEN Ai-min(National-Local Joint Engineering Laboratory of Geo-spatial Information Technology,Hunan University of Science and Technology,Xiangtan 411201;Hunan Province Engineering Laboratory of Geospatial Information,Hunan University of Science and Technology,Xiangtan 411201;Xiangtan Geospatial Information Application Engineering Technology Center,Xiangtan 411201,China)
出处
《地理与地理信息科学》
CSCD
北大核心
2018年第4期13-18,共6页
Geography and Geo-Information Science
基金
国家自然科学基金项目(41571374)
湖南省重点实验室开放基金项目(JL16K01
CT16K02)
湖南省教育厅重点项目(16A070)
湖南省自然科学基金湘潭联合基金项目(2017JJ4037)
关键词
视觉认知
三维建筑群模型
几何阈值
层次概括
聚类概括
visual cognition
3D building group model
geometric threshold
hierarchical generalization
clustering generalization
