A Clustering Algorithm of Electronic Boundary Points for Cadastral Blocks

The coordinates of boundary points of a number of cadastral blocks can be obtained by electronic boundary points automatically, while the automation of data processing of coordinates needs to be improved. This paper focuses on the automatic clustering of electronic boundary point coordinates and analyses the location distribution and angle characteristics of cadastral blocks with real coordinates data from 6 villages. In this paper, a strategy is proposed that one boundary point of every cadastral block should be marked and Manhattan distance between points should be used as the similarity measure. A clustering algorithm of electronic boundary points for cadastral blocks is designed with the combination of K-means algorithm. The simulation results show the clustering accuracy of the algorithm can be above 0.9 when the number of cadastral blocks is more than 100, and it can be improved as the number of cadastral blocks is controlled within 25. This study provides a reference for optimization of existing electronic boundary points system and automation technology of the cadastral survey.