Multi-aspect techniques in remote sensing

Abstract

Multi-aspect techniques, which use multispectral scanner information collected from two or more different view angles for each spatial element, are discussed. Empirical recognition results obtained with various combinations of information channels are presented. Theoretical calculations of multi-aspect reflectances of corn fields also are analyzed and compared with empirical data. Also presented is a brief comparison between recognition processing and scene attribute estimation techniques for extracting information from multispectral scanner data. One method for improving information extraction by employing a post-recognition attribute estimation procedure is suggested.