An Operational Cloud Classifier Based On A Self-Organized Feature Map
Kimmo Valkealahti, Jukka Iivarinen, Ari Visa, Olli Simula
Helsinki University of Technology
Laboratory of Information and Computer Science
Rakentajanaukio 2 C, SF-02150 Espoo
Finland
Abstract
Experiences from a new method to interpret satellite images is represented.
The described method is adaptive. It is trained by examples. In the reported
application a combination of texturel and spectral measures is used as
a feature vector. The adaptation of the extracted feature vectors occurs
by a self-organizing process. As a result a topological feature map is
generated. The map is identified and fine tuned by known samples, examples
of clouds. The map is used later on as a codebook for cloud classification.
The obtained verification results are good. The represented method is general
in the sense that with new features the method can be applied to new problems.
Report A19, December 1993
Sorry, no postscript version available!
Back to Cloud Project Home Page