Geometric variability of raster cell class assignment

Abstract

Applications like land-cover mapping frequently require the transformation of vector polygons to raster grids, such that each cell in the output raster is assigned a single class. Cell value assignment is a direct function of the geometric relationship between the vector polygons and the origin, orientation and spatial resolution of the raster structure, acting in conjunction with the value assignment rule employed by the algorithm. Results of an alternative approach to characterizing origin assignment variability are reported. A simulation approach is employed to perturb the origin of the grid lattice by amounts smaller than the grid spacing. The distribution of results across all output raster realizations indicates the rasterization variability for a particular input vector map at a particular cell size. Results for a representative vector land-cover dataset are presented, indicating that substantial variability is possible. The paper concludes with a discussion of theoretical and practical implications for the identification of appropriate cell resolutions, as well as for identifying particular input shapes and polygon configurations that may be especially sensitive to rasterization error.