Areal interpolation between one partitioning of geographical space and another remains an important topic, particular in terms of population counts and related statistics which are often required in order to compute an incidence ratio. Despite numerous recent developments in intelligent areal interpolation methods, and studies that have demonstrated their clear advantage over simple areal weighting, there is little evidence to suggest widespread usage amongst the GIS user community. It is argued that to encourage greater uptake such methods must offer simplicity and convenience. Areal interpolation based on binary dasymetric mapping is conceptually simple, but examples to date tend to use information extracted from multi-spectral satellite imagery which limits its perceived convenience. This paper examines a simple method to extract equivalent information from a raster pixel map. It is shown to offer comparable areal interpolation performance at considerably less cost in terms of both time and complexity.
|19 - 32
|Number of pages
|Computers Environment and Urban Systems
|Published - 2007