A Composite Influence Domain Model for Automatically Selecting Islands in Nautical Charts

Abstract

Existing methods for automatically selecting islands for nautical charts utilize the Voronoi diagram. However, we will show that the Voronoi diagrams cannot accurately represent the density and distribution of islands in cartographic generalization, which makes it difficult to obtain selection results that meet requirements. We propose a novel method for selecting islands based on influence domains. First, the shortcomings of the Voronoi diagram representation of spatial data are analyzed using an island influence domain (IID) model. Second, the model is defined and constructed, and the importance of an island is weighed according to its attributes. Third, several islands are selected automatically by choosing those with the largest area of IID and deleting those whose IIDs overlapped more with pre-selected islands. Finally, several groups of islands with different types of sea areas are selected for experimentation, and the required selection parameters are determined and analyzed to derive an empirical formula for setting the parameters. The experimental results show that: (1) the proposed method improves the quality of island selection; (2) the empirical formula can be used in parameter setting to obtain acceptable results.