Geostatistical Mapping of Satellite Data using P-field Simulation with Conditional Probability Fields

摘要

This paper presents a variant of p-field simulation that allows the generation of spatial realizations through the sampling of a set of conditional probability distribution functions by conditional probability fields. The approach is illustrated using a randomly sampled (200 observations of the NIR channel) SPOT scene of a semi-deciduous tropical forest. Results indicate that the use of conditional probability fields improves the reproduction of statistics such as histogram and semivariogram, while yielding more accurate predictions of reflectance values than the common p-field implementation or the more CPU-intensive sequential indicator simulation. The proposed approach also leads to a better prediction of the size of contiguous areas covered by savannah.