Stochastic Signal Processing > Variograms

In a probability model, the theoretical variogram is defined as the variance of the variable Z of x plus h minus Z of x. It quantifies the spatial variability of the random variable Z of x as a function of the distance h. Under the stationary assumption, there is a link between the variogram and covariance.

As the covariance expresses the spatial correlation, the variogram expresses the spatial variability.

In signal processing, the variogram is equivalent to the autocorrelation function. It is the inverse Fourier transform of the frequency power spectrum.