Multispectral imagery has been used by geoscientific communities for more than 25 years. Problems have resulted, however, in what appears to be a growing disjoint between digital imagery processing and traditional interpretation. Mainly, the problems relate to use of algorithms -- from which there are thousands to choose -- without proper consideration of interprtation foundations. The unfortunate results suggest that current instructional efforts, as dynamic and sophisticated as they may be, are geared toward processing techniques but are in absence of interpretive foundations. Current trends suggest this may be changing, but more work needs to be devoted to joining processing and interpretation in the context of data utilization. This is particularly true when one considers the products that are emerging as industry 'standards' whereby multispectral scientists must have firm understandings of linear analytical strategies if their eventual outputs are to have sufficient reliabilities and validities.

Date Published: 1 September 1995
PDF: 7 pages
Proc. SPIE 2555, Airborne Reconnaissance XIX, (1 September 1995); doi: 10.1117/12.218604

Published in SPIE Proceedings Vol. 2555:
Airborne Reconnaissance XIX
Wallace G. Fishell; Arthur A. Andraitis; Paul A. Henkel; Alfred C. Crane Jr., Editor(s)