Modern G. I. S. technology permits extensive processing of geographical data oriented to geometrical treatment of space. In most cases, digital information is available in a topological data format which is not the one required. Then data transformations from one topology to another have to be performed, and consequently a bias is introduced in the results of data analysis. Because of this problem with spatial data handling and analysis, planners and regional analysts require methods capable of transferring attribute data from one given topology to another. The approach we propose is to define spatial data transformations through linear operators applied to original attribute variables, so as to produce derived maps with a different topology. Within this framework, the operation of making inferences on the original data can be viewed as an inverse linear transformation problem. Assuming a priori that local map characteristics follow a Gaussian Markov random field, this problem can be formalized through Bayes' theorem and solved by the classical maximum a posteriori estimation procedure. © 1998 OPA (Overseas Publishers Association) N.V. Published by license under the Gordon and Breach Science Publishers imprint.
|Pages (from-to)||199 - 220|
|Number of pages||22|
|Publication status||Published - 1998|
All Science Journal Classification (ASJC) codes
- Earth and Planetary Sciences (miscellaneous)