LANDSAT-TM image classification using principal components analysis and neural networks

The application of three neural architectures in the classification LANDSAT images is conducted using multispectral data as well as the principal components projections. Results evaluation is given in terms of recognition rates and reconstructed images