PRINCIPAL COMPONENTS TRANSFORM WITH SIMPLE, AUTOMATIC NOISE ADJUSTMENT

A new form of the Principal Components transform is described which is particularly suited for use with hyperspectral image data, such as th e images produced by the Airborne Visible/Infrared Imaging Spectromete r (AVIRIS). This new transform scales or adjusts the image data in eac h band by an estimate of the noise in each band. The noise estimates a re simply made from the image data itself through the inverse of its c ovariance matrix. For reasons associated with this, the transform is c alled a 'Residual-scaled' PC or RPC transform. The inversion of the co variance matrix is the only extra computation required over and above that needed for the ordinary PC transform. The RPC transform correspon ds to using a diagonal noise matrix with the Maximum Noise Fraction tr ansform or the Noise-Adjusted PC transform. Its performance is compare d with that of the ordinary PC and the Standardized PC transforms for 102 bands of a 1992 AVIRIS image of a vegetated area (the Jasper Ridge Biological Preserve). Its low-order, high-variance components are of consistently better quality than theirs. The Standardized PC transform performs poorly with such hyperspectral data and should be used with caution, if at all.