MIXTURE-MODELS WITH HIGHER-ORDER MOMENTS

In this paper we present a novel method for mixed pixel classification where the classification of groups of mixed pixels is achieved by tak ing into consideration the higher order moments of the distributions o f the pure and the mixed classes. The equations expressing the relatio nship between the higher order moments are used to augment the set of equations that express the relationship between the means only. We sho w that weighting these equations does not make the set of equations av ailable less reliable. As a consequence, the number of equations can b e increased and thus more classes than available bands can be identifi ed. The method is exhaustively tested using simulated data and is also applied to real Landsat TM data for which ground data are available.