n this paper the potential of neural networks has been applied to hyperspectral data and exploited either for classification purposes or for data feature extraction and dimensionality reduction. For this latter task, a topology named autoassociative neural network has been used. In its complete form, the processing scheme uses a neural network architecture consisting of two stages: the first stage reduces the dimension of the input vector while the second stage performs the mapping from the reduced input vector into the land cover classification. The effectiveness of the feature extraction algorithm has been evaluated for a set of experimental data provided by the AHS radiometer comparing its performance to that obtainable with more traditional linear techniques such as PCA, while the accuracy of the final classification map has been computed on the base of the available ground-truth
DEL FRATE, F., Licciardi, G., Duca, R. (2009). Autoassociative neural networks for features reduction of hyperspectral data. In Proceedings of the First Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing. IEEE [10.1109/WHISPERS.2009.5288997].
Autoassociative neural networks for features reduction of hyperspectral data
DEL FRATE, FABIO;
2009-01-01
Abstract
n this paper the potential of neural networks has been applied to hyperspectral data and exploited either for classification purposes or for data feature extraction and dimensionality reduction. For this latter task, a topology named autoassociative neural network has been used. In its complete form, the processing scheme uses a neural network architecture consisting of two stages: the first stage reduces the dimension of the input vector while the second stage performs the mapping from the reduced input vector into the land cover classification. The effectiveness of the feature extraction algorithm has been evaluated for a set of experimental data provided by the AHS radiometer comparing its performance to that obtainable with more traditional linear techniques such as PCA, while the accuracy of the final classification map has been computed on the base of the available ground-truthI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.