A new per-field classification method using mixture discriminant analysis

Abstract

In this study, a new per-field classification method is proposed for supervised classification of remotelysensed multispectral image data of an agricultural area using Gaussian mixture discriminant analysis(MDA). For the proposed per-field classification method, multivariate Gaussian mixture models constructedfor control and test fields can have fixed or different number of components and each component can havedifferent or common covariance matrix structure. The discrimination function and the decision rule of thismethod are established according to the average Bhattacharyya distance and the minimum values of theaverage Bhattacharyya distances, respectively. The proposed per-field classification method is analyzedfor different structures of a covariance matrix with fixed and different number of components. Also, weclassify the remotely sensed multispectral image data using the per-pixel classification method based onGaussian MDA.