Glossary for Geospatial Science

cubic convolution or cubic interpolation:  A computationally intense interpolation method used in raster resampling.  This method determines a new cell value to produce a smoother result by fitting a cubic polynomial surface to a 4 x 4 neighborhood of cells.  Bilinear interpolation normally produces results which are almost as good.  Cubic convolution should not be applied where the mean data value for a group of cells is undefined as in categorical data (like an 8-bit color raster object, a classification map, or any raster object containing similar category-type data).