It is often useful to estimate obscured or missing remotely sensed dat
a. Traditional interpolation methods, such as nearest-neighbor or bili
near resampling, do not take full advantage of the spatial information
in the image. An alternative method, a geostatistical technique known
as indicator kriging, is described and demonstrated using a Landsat T
hematic Mapper image in southern Chiapas, Mexico. The image was first
classified into pasture and nonpasture land cover. For each pixel that
was obscured by cloud or cloud shadow, the probability that it was pa
sture was assigned by the algorithm. An exponential omnidirectional va
riogram model was used to characterize the spatial continuity of the i
mage for use in the kriging algorithm. Assuming a cutoff probability l
evel of 50%, the error was shown to be 17% with no obvious spatial bia
s but with some tendency to categorize nonpasture as pasture (overesti
mation). While this is a promising result, the methods practical appli
cation in other missing data problems for remotely sensed images will
depend on the amount and spatial pattern of the unobscured pixels and
missing pixels and the success of the spatial continuity model used.