Hyperspectral data analysis using wavelet-based classifiers

In general, the analysis of hyperspectral remote sensing data by means of pattern recognition and/or classification is known to be data dependent. Thus, conventional methods for classifications may not be applicable due to the large amount of data collection used to characterize hyperspectral data in terms of optimality and computational time. In this paper, wavelet-based classifiers are presented and hyperspectral signatures are extracted from the available data and then used for the discrimination of various sample types of vegetation.