摘要
建筑群空间分布模式识别对制图综合、多尺度表达及空间数据挖掘具有重要意义。针对建筑群中以建筑物组合结构为单元的直线模式识别问题,提出一种建筑群同质二元组直线模式的识别方法。首先分析研究同质二元组直线模式的认知特征和定义;然后利用Delaunay三角网构建建筑群邻近关系,以建筑物邻近性、尺寸和方向相似性约束进行聚类,考虑邻近程度的异质性、直线性识别直线模式,根据直线模式识别结果识别同质二元组结构;最后逐步识别由具有相似特征且呈等间隔直线排列的同质二元组构成的直线模式。实验表明所提方法能有效识别出同质二元组直线模式,其识别结果符合人类认知。
Spatial distribution pattern recognition of buildings is significant to cartographic generalization,multi-scale representation and spatial data mining.In this paper,an approach is presented to recognize homogeneous two-tuple collinear patterns with the units which are the structures of buildings combination.Firstly the cognitive characteristics and definitions of homogeneous two-tuple collinear pattern are analyzed.Then the neighborhood relationship is captured by Delaunay triangulation and clustering is carried out with constraints of building adjacency,size and orientation similarity.Considering the heterogeneity of adjacency and straightness,the homogeneous two-tuple structure is identified according to the result of linear pattern recognition.Finally,collinear patterns composed homogeneous two-tuple with similar features and arranged in straight lines at equal intervals are gradually identified.Experiments show that the proposed method is effective for homogeneous two-tuple collinear pattern recognition with the agreement of human spatial cognitive characteristics.
作者
行瑞星
武芳
巩现勇
杜佳威
刘呈熠
殷吉崇
XING Ruixing;WU Fang;GONG Xianyong;DU Jiawei;LIU Chengyi;YIN Jichong(Information Engineering University, Zhengzhou 450001, China)
出处
《测绘科学技术学报》
CSCD
北大核心
2021年第1期91-96,共6页
Journal of Geomatics Science and Technology
基金
国家自然科学基金项目(41471386,41801396)。
关键词
制图综合
建筑群
视觉感知理论
同质二元组直线模式
模式识别
cartographic generalization
building groups
visual perception theory
homogeneous two-tuple collinear pattern
pattern recognition