Fuzzy hierarchical analysis for remotely sensed data

Fuzzy interpretations of data structures are a very natural and intuitively plausible way to formulate and solve various problems such as uncertainty, vagueness, decision making etc. A nonparametric hierarchical fuzzy classification for satellite data is proposed. Incorporation of symbolic data analysis to multispectral data is helpful in data reduction and feasible to represent the complex description. Hence, remotely sensed data are converted into symbolic form. Subjective similarity to each pair of symbolic objects is computed to perform classification. The proposed algorithm has two stages. In the first stage, multispectral data are converted into symbolic form. The second stage involves classification of symbolic objects using a fuzzy equivalence relation. The equivalence relation is found using transitive closure for the distance computed between a pair of symbolic objects. Distance is computed with a modified distance measure, which handles the quantitative interval type of data. The classification results of the IRS-1B satellite data covering Hyderabad city, India are encouraging. The results confirm that the algorithm is efficient with less computational time, less memory and able to classify sharp boundaries