Nested Spatial Models for Image Analysis Applications

Keywords: variograms, nested models, kriging, cokriging

Abstract: Variogram analysis, modeling and estimation is a common approach to analysis of spatial data. Variogram models traditionally have been simple. For some spatial data (such as digital imagery) a double or nested model provides a better fit than simpler models. Such models offer benefits in such areas as filtering, sampling, and compression. These models may also prove useful in multivariate and other nonlinear spatial estimation models. Applications of this approach to real data from multispectral imagery will be demonstrated.