Acta Geodaetica et Cartographica Sinica ›› 2021, Vol. 50 ›› Issue (6): 800-811.doi: 10.11947/j.AGCS.2021.20200298

• Cartography and Geoinformation • Previous Articles     Next Articles

The template matching approach to combined collinear pattern recognition in building groups

XING Ruixing, WU Fang, GONG Xianyong, DU Jiawei, LIU Chengyi   

  1. Institute of Geospatial Information, Information Engineering University, Zhengzhou 450001, China
  • Received:2020-07-07 Revised:2021-04-20 Published:2021-06-28
  • Supported by:
    The National Natural Science Foundation of China (Nos. 41471386;41801396)

Abstract: Spatial distribution pattern recognition of buildings is significant to cartographic generalization, multi-scale representation and spatial data mining. This paper presents an approach to recognize combined collinear patterns with local heterogeneity. Firstly the cognitive characteristics and definitions of combined collinear pattern are analyzed. Then neighborhood relationship and extension alignment between buildings are obtained by clustering based on the proximity, size and orientation. Take the enlarged building’s smallest minimum bounding rectangle as the initial matching template. Finally, considering the constraints of straightness, similarity and local heterogeneity, the combined linear patterns are extracted by searching and matching buildings through continuous templates constructed based on distribution spacing and direction. Experiments show that the proposed method is effective for combined collinear pattern recognition with the agreement of human spatial cognitive characteristics.

Key words: cartographic generalization, building groups, visual perception theory, combined collinear pattern, pattern recognition

CLC Number: