Multispectral Classification of Landsat-Images using Neural Networks (bibtex)
by Horst Bischof, Werner Schneider, Axel Pinz
Abstract:
Recent progress in neural network research has demonstrated the usefulness of neural networks in a variety of areas. In this work we report the application of three-layer backpropagation networks for classification of Landsat TM data on a pixel by pixel basis. The results are compared to Gaussian maximum likelihood classification. It is shown that the neural network is able to perform better than the maximum likelihood classifier. In an extension of the basic network architecture it is shown that textural information can be integrated into the neural network classifier without the explicit definition of a texture measure. The usage of neural networks for postclassificational smoothing is examined.
Reference:
Multispectral Classification of Landsat-Images using Neural Networks (Horst Bischof, Werner Schneider, Axel Pinz), Technical report, PRIP, TU Wien, 1991.
Bibtex Entry:
@TechReport{TR004,
  author =	 "Horst {Bischof} and Werner {Schneider} and Axel
                  {Pinz}",
  institution =	 "PRIP, TU Wien",
  number =	 "PRIP-TR-004",
  title =	 "{M}ultispectral {C}lassification of
                  {L}andsat-{I}mages using {N}eural {N}etworks",
  year =	 "1991",
  url =		 "ftp://ftp.prip.tuwien.ac.at/pub/publications/trs/tr4.ps.gz",
  abstract =	 "Recent progress in neural network research has
                  demonstrated the usefulness of neural networks in a
                  variety of areas. In this work we report the
                  application of three-layer backpropagation networks
                  for classification of Landsat TM data on a pixel by
                  pixel basis. The results are compared to Gaussian
                  maximum likelihood classification. It is shown that
                  the neural network is able to perform better than
                  the maximum likelihood classifier. In an extension
                  of the basic network architecture it is shown that
                  textural information can be integrated into the
                  neural network classifier without the explicit
                  definition of a texture measure. The usage of neural
                  networks for postclassificational smoothing is
                  examined.",
}
Powered by bibtexbrowser