Quality assessment in point feature generalization with pattern preserved
Wenhao Yu · Transactions in GIS · 2018
Abstract Geographical features often show certain spatial patterns on a map in terms of the arrangement of point symbols. These patterns are essentially related to the underlying geographical processes and landscapes. Thus, when deriving small‐scale maps from a large‐scale map, one of the most important constraints that cartographers or systems should follow is to retain the basic patterns of point objects on the target map. However, no research in the literature currently evaluates the quality of point feature generalization in terms of spatial pattern. This study proposes an approach to quantitatively measure the pattern change after generalization. The basic idea of the approach is to extend advanced image analysis techniques (e.g., texture recognition) to measure the patterns of point objects in a map space. Specifically, there are two main steps: firstly, the original space is converted into the raster space by utilizing a regularly spaced grid (i.e., a grayscale image) with cell attributes representing the local intensity level of point features; secondly, the texture analysis operation is performed on the grid to obtain the feature descriptors of the point pattern. The experimental results demonstrate that the proposed approach is effective in comparing the point patterns before and after generalization.