COMPARISON OF FUZZY C-MEANS CLASSIFICATION, LINEAR MIXTURE MODELING AND MLC PROBABILITIES AS TOOLS FOR UNMIXING COARSE PIXELS

Three different 'soft' classifiers (fuzzy c-means classifier, linear m ixture model, and probability values from a maximum likelihood classif ication) were used for unmixing of coarse pixel signatures to identify four land cover classes (i.e., supervised classifications). The coars e images were generated from a 30 m Thematic Mapper (TM) image; one se t by mean filtering, and another using an asymmetric filter kernel to simulate Multi-Spectral Scanner (MSS) sensor sampling. These filters c ollapsed together windows of up to 11 x 11 pixels. The fractional maps generated by the three classifiers were compared to truth maps at the corresponding scales, and to the results of a hard maximum likelihood classification. Overall, the fuzzy c-means classifier gave the best p redictions of sub-pixel landcover areas, followed by the linear mixtur e model. The probabilities differed little from the hard classificatio n, suggesting that the clusters should be modelled more loosely. This paper demonstrates successful methods for use and comparison of the cl assifiers that should ideally be extended to a real dataset.