Open    Access



Constrained Delaunay Triangulation for Grouping Functional Areas by Land Use



Author(s)

Jirong Gu, Xianwei Cheng and Zhi Dou

Full-Text PDF XML     674 Views

DOI:10.17265/2328-2193/2016.04.001

Functional land use maps are used for land evaluation, environmental analysis, and resource conservation. Spatial data clustering identifies the sparse and crowded places, thus discovering the overall distribution pattern of the dataset. Some clustering methods represent an attribute-oriented approach to knowledge discovery. Other methods rely on natural notions of similarities (e.g., Euclidean distances). These are not appropriate for constructing functional areas. We propose a similarity value to evaluate the closeness between a pair of points based on the total functional area and the proportion of the main land use type for the entire functional area. We develop constrained attributes employing this similarity value and a DT (Delaunay triangulation) criterion function when merging clusters. Four thresholds are set to ensure that functional areas have acceptable proportions, regular shapes, and no overlap. An experimental study was conducted with cadastral data for Chengdu, China, from 2009. The results show the advantages for objectivity and efficiency in using the proposed algorithm to define functional areas. The areas are created dynamically at any convenient time.

Constrained Delaunay triangulation, functional areas, land use, similarity, constrained clustering.