Hyperspectral Image Classification Using Gaussian Mixture Models and Markov Random Fields

• Published in: IEEE geoscience and remote sensing letters Vol. 11; no. 1; pp. 153 - 157
• Main Authors: Li, Wei, Prasad, Saurabh, Fowler, James E.
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2014; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Classification; Dimensionality reduction; Gaussian mixture model; Gaussian mixture model (GMM); hyperspectral classification; Hyperspectral imaging; Markov processes; Markov random field (MRF); nonnegative matrix factorization; Training
• ISSN: 1545-598X; 1558-0571
• DOI: 10.1109/LGRS.2013.2250905

The Gaussian mixture model is a well-known classification tool that captures non-Gaussian statistics of multivariate data. However, the impractically large size of the resulting parameter space has hindered widespread adoption of Gaussian mixture models for hyperspectral imagery. To counter this parameter-space issue, dimensionality reduction targeting the preservation of multimodal structures is proposed. Specifically, locality-preserving nonnegative matrix factorization, as well as local Fisher's discriminant analysis, is deployed as preprocessing to reduce the dimensionality of data for the Gaussian-mixture-model classifier, while preserving multimodal structures within the data. In addition, the pixel-wise classification results from the Gaussian mixture model are combined with spatial-context information resulting from a Markov random field. Experimental results demonstrate that the proposed classification system significantly outperforms other approaches even under limited training data.