Title :
An incremental learning classifier for remote-sensing images
Author :
Burzzone, L. ; Prieto, Diego Fernàndez ; Cossu, Roberto
Author_Institution :
Dept. of Biophys. & Electron. Eng., Genoa Univ., Italy
Abstract :
An incremental learning classifier based on RBF neural networks is presented. The proposed classifier relies on an incremental learning technique that allows the periodical acquisition of new knowledge when new training sets became available, without suffering a significant loss of the existing one. Such a characteristic renders the presented classifier a promising tool for the monitoring of extended geographical areas in a regular basis
Keywords :
geophysical signal processing; geophysical techniques; geophysics computing; image classification; knowledge engineering; learning (artificial intelligence); radial basis function networks; remote sensing; terrain mapping; RBF neural network; geophysical measurement technique; image classification; imaging; incremental learning classifier; knowledge acquisition; land surface; neural net; neural network; periodical acquisition; radial basis function; remote sensing; remote-sensing image; terrain mapping; training set; Computer architecture; Costs; Electronic mail; Kernel; Neural networks; Neurons; Remote monitoring; Remote sensing; Rendering (computer graphics); Robustness;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 1999. IGARSS '99 Proceedings. IEEE 1999 International
Conference_Location :
Hamburg
Print_ISBN :
0-7803-5207-6
DOI :
10.1109/IGARSS.1999.771550
