Profile Retrieval Estimation Techniques

Abstract

Retrieval techniques, applicable to problems where measurements have to be blended, satisfy an equation that can be approximated by a Fredholm integral equation of the first kind. A variety of inversion techniques, that have been used in radiometric remote sensing of temperature and moisture profiles, are outlined to solve this equation focusing on the wide variety of supplementary information that may be used in addition to radiance information. Techniques, such as regularization and iterative methods, are introduced because of their simplicity and explicit use of an initial guess. Methods based on Backus-Gilbert and Kalman approaches are shown for diagnosis problems. Statistical techniques are illustrated emphasizing the use a priori information to constrain the solution. Final remarks give the current scenario on this subject.