Semi-Supervised Support Vector Biophysical Parameter Estimation

Author

Camps-Valls, G. ; Munoz-Mari, J. ; Gomez-Chova, Luis ; Calpe-Maravilla, J.

Author_Institution

Dept. Eng. Electron., Univ. of Valencia, Valencia

Volume

3

fYear

2008

fDate

7-11 July 2008

Abstract

Two kernel-based methods for semi-supervised regression are presented. The methods rely on building a graph or hypergraph Laplacian with both the labeled and unlabeled data, which is further used to deform the training kernel matrix. The deformed kernel is then used for support vector regression (SVR). The semi-supervised SVR methods are sucessfully tested in LAI estimation and ocean chlorophyll concentration prediction from remotely sensed images.

Keywords

geophysical signal processing; geophysical techniques; image processing; oceanography; regression analysis; remote sensing; support vector machines; vegetation; LAI estimation; biophysical parameter estimation; hypergraph Laplacian building; kernel based methods; ocean chlorophyll concentration prediction; remotely sensed images; semisupervised support vector regression; training kernel matrix deformation; Condition monitoring; Kernel; Laplace equations; Neural networks; Noise robustness; Oceans; Parameter estimation; Remote monitoring; Spatial resolution; Testing;

fLanguage

English

Publisher

ieee

Conference_Titel

Geoscience and Remote Sensing Symposium, 2008. IGARSS 2008. IEEE International

Conference_Location

Boston, MA

Print_ISBN

978-1-4244-2807-6

Type

conf

DOI

10.1109/IGARSS.2008.4779554

Filename

4779554