Evolution of Low Complexity Artificial Neural Networks for Land Cover Classification from Remote Sensing Data
Artificial Neural Networks (ANN) have gained increasing popularity as
an alternative to statistical methods for classification of Remote
Sensing Data. Their superiority to some of the classical statistical
methods has been shown in the literature. Therefore, ANNs are commonly used
for segmentation and classification purposes. We address the problem of generating an appropriate low complexity network topology, the right number of training epochs and preprocessing the training data set for multi-layer feed-forward ANNs. A method based on Genetic Algorithms (GA) for the automatic generation of problem-adapted topologies is employed with the parallel netGEN system which has been designed by the authors. A land cover classification problem using multi-spectral Landsat Thematic Mapper (TM) data is presented so as to demonstrate the capabilities of netGEN.
Helmut A. Mayer
<helmut@cosy.sbg.ac.at>
Last modified: Mon Jan 5 1998