Gestalt-Based Douglas-Peucker Algorithm to Keep Shape Similarity and Area Consistency of Polygons

Abstract

Douglas-Peucker (DP for short) algorithm plays an important role in vector data simplification and map generalization field. However, it is difficult to keep the similarity between generalized data and original data, especially when an inappropriate tolerance is set for the DP algorithm. What's more, area of the simplified polygon by DP algorithm would change a lot in some specific tolerance distance cases. The paper proposed a Gestalt based Douglas-Peucker algorithm (GDP algorithm for short). The GDP algorithm took into account spatial distribution characteristics of all vertexes and adjusted some points generated by DP algorithm. It could keep similarity of vector data and made polygon area unchangeable. Finally, the paper applied GDP and DP algorithm with different tolerances in experiments of line generalization and polygon generalization. The results of Experiment showed that GDP algorithm is more stable and keeps higher data quality of vector data than DP algorithm. In addition, GDP algorithm can ensure that the area of polygon does not change after simplification.