De-noising reflectance spectra: Simulation experiments and applications in land degradation studies

摘要

The technique of derivative analysis, developed initially in the field of analytical chemistry, has been applied in the analysis of high spectral resolution data. However, it is widely known that derivative analysis is sensitive to noise in the data; hence, effective noise removal is needed if derivative analysis is to be applied successfully. Various methods of noise removal are described in the literature. A newly developed method based on the wavelet transform appears promising, though there is little practical guidance on its use. In this study, the investigation of several important parameters that govern wavelet-based de-noising is undertaken. The optimal parameter settings are then evaluated for use in derivative analysis using field spectroradi-ometer data and in mapping vegetation cover for land degradation studies from DAIS 7915 airborne imaging spectrometry (IS) data using spectral red edge parameters. Results indicated that wavelet-based de-noising is useful in reducing noise and thus facilitating derivative analysis, though some limitations were encountered.