A methodology for balancing the preservation of area, shape, and topological properties in polygon-to-raster conversion

ABSTRACT

Polygon-to-raster conversion inevitably introduces a loss in spatial properties of polygons, such as area or topology, which should be preserved. Existing methods preserve only one property, resulting in greater losses in other properties. In this study, we propose a new methodology to balance the preservation of area, shape, and topological properties during conversion. By reassigning cells of the rasterized outcome, the method first compensates for the loss in shape properties. Topological changes are then corrected by comparing the topological relations of raster regions and their corresponding polygons. Finally, the areas between pairs of neighboring regions are coordinated to maintain area properties. The main contribution of this study relies on the fact that the presented method considers the interactions of different properties, rather than separately preserving each of them. We employed a land-use dataset containing 14,000 polygons for our experiments. When the cell size increased from 5 to 25 m, the presented method resolved 48.4% of overall rasterization errors on average, which was much higher than those of the area-, shape-, and topology-preserving methods (i.e. 2.6%, 26.7%, and 34./0%, respectively). However, the presented method increased the computational time by 579%, 264%, and 52%, respectively, as compared with these three methods.

KEYWORDS:

• Area
• polygon-to-raster conversion
• preservation
• rasterization errors
• shape
• topology

Acknowledgments

We sincerely thank the editors and anonymous reviewers for their constructive comments, which have helped improve this paper.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data Availability Statement

The computer code and sample dataset that support the findings of this study are available at https://doi.org/10.6084/m9.figshare.14227028. The code was developed using Python 3.8 and released in March 2021. It is recommended to run the code on GDAL 3.2 or later.

Additional information

Funding

This work was supported by the National Key Research & Development Program of China (grant number 2017YFB0504205) and the National Natural Science Foundation of China (grant number 41901318).