Tor Vergata Earth Observation Laboratory |
A user-friendly automatic tool for image classification based on neural networks B. Buttarazzi, F. Del Frate, C. Solimini
In this paper the implementation and the functionality of software created on the basis of neural networks technology is explained. The program developed allows the user to read an image, to create and to train a neural network, and finally, to classify the entire image. Using the implemented software it is possible to develop a unique program which processes efficiently all the data. Actually, one of the thematic principals of this material is the classification of satellite images. Given that an image acquired by the satellite could be composed by different bands not visible to the naked eye (and therefore not interpretable), an automatic program, automatically trained, is the ideal method to study the various of information contained in it. INTERNATIONAL SOCIETY OF PHOTOGRAMMETRY AND REMOTE SENSING Proceedings of the ISPRS joint conference 5th International Symposium Remote Sensing of Urban Areas (URS 2005) Tempe, AZ, USA, March 14-16 2005 Editors: M. Moeller, E. Wentz International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences © 2005 the authors
|
|| full paper
|| earth observation laboratory || |