A Topology-Preserving Polygon Rasterization Algorithm

Chen Zhou,Dingmou Li,Ningchuan Xiao,Zhenjie Chen,Xiang Li,Manchun Li

DOI: https://doi.org/10.1080/15230406.2017.1401488

IF: 2.354

2017-01-01

Cartography and Geographic Information Science

Abstract:Conventional algorithms for polygon rasterization are typically designed to maintain non-topological characteristics. Consequently, topological relationships, such as the adjacency between polygons, may also be lost or altered, creating topological errors. This paper proposes a topology-preserving polygon rasterization algorithm to avoid topological errors. Four types of topological error may occur during polygon rasterization. The algorithm starts from an initial polygon rasterization and uses a set of preserving strategies to increase topological accuracy. The count of the four types of error measures the topological errors of the conversion. Topological accuracy is summarized as 1 minus the ratio of actual topological errors to the total number of possible error cases. When applied to a land-use dataset with a data volume of 128 MB, 127,836 polygons, and extending 1352 km(2), the algorithm achieves a topological accuracy of more than 99% when raster cell size is 30 m or smaller (100% for 5 and 10 m). The effects of cell size, polygon shape, and number of iterations on topological accuracy are also examined.