RAMM/CIRA Quarterly Report

Principal Component Images (PCIs) have been generated from over 200 visible/near-IR bands of EO-1 Hyperion data. The hyper-spectral data were converted to AREA format to accommodate PCI software developed for McIDAS. Initial reactions are that PCIs are slow to compute on large numbers of bands, and that PCIs are highly susceptible to the large explained variance due to noise in the Hyperion data. Solutions to these problems are being explored.

Two manuscripts on analysis of MODIS imagery using Principal Component Image analysis have been accepted by the Journal of Applied Meteorology.