Application of simulated annealing in optical remote sensing of marine constituents

Oyvind Frette, Jakob J. Stamnes, Knut Stamnes, Svein Rune Erga

Research output: Contribution to journalConference articlepeer-review

1 Scopus citations

Abstract

Optical remote sensing of ocean colour is a well-established technique. But most algorithms developed hitherto have been based on the assumption that only the phytoplankton affect the optical properties of the ocean. Such algorithms are often based on assumptions that become questionable in coastal areas. The assumption of a near-infrared dark pixel in the satellite image, will no longer be valid, and the band-ratio technique used for computing the algae concentration will also become inaccurate. To overcome these limitations we have developed an inverse-modelling algorithm for retrieval of marine constituents. Here the determination of ocean colour is based on a three-component optical model consisting of chlorophyll-a, suspended matter, and yellow substance. We also use one parameter to describe the thickness of the aerosol layer. A simulated-annealing optimization scheme is employed to minimize the difference between measured satellite data and corresponding simulated data obtained using a coupled atmosphere-ocean radiative transfer code. The same optimization method has also been applied to the problem of retrieving the algae concentration in waters with vertical structure. In this case the marine parameters of interest are the algae concentration in two different layers as well as the thickness of the first layer.

Original languageEnglish
Pages (from-to)262-270
Number of pages9
JournalProceedings of SPIE - The International Society for Optical Engineering
Volume3821
StatePublished - 1999
EventProceedings of the 1999 Environmental Sensing and Applications - Munich, Ger
Duration: 14 Jun 199917 Jun 1999

Fingerprint

Dive into the research topics of 'Application of simulated annealing in optical remote sensing of marine constituents'. Together they form a unique fingerprint.

Cite this