Carlos Alberto Felgueiras,  Suzana Druck Fuks,  Antônio Miguel Vieira Monteiro,  Eduardo Celso Gerbi Camargo. 

Raster representations of thematic and numerical spatial attributes are very common in a GIS environment for computational simulation and analysis of spatial processes. This paper addresses the problem of predictions with uncertainty assessment for GIS raster representations created from a set of sample points of spatial attributes. The realizations of a stochastic simulation inference process, over numerical attribute samples, are used in order to infer the attribute values and related uncertainties at non-sampled spatial locations. A case study, using elevation sample data, is presented to illustrate the concepts used in this work.

http://www.tecgraf.puc-rio.br/geoinfo2000/anais/017.pdf

Processando consulta.