Redes neuronales artificiales para la clasificación de imágenes satelitales

Abstract

In this paper is made the analysis to data set thatrepresent different images, classified this way: redearth, cotton crop, gray earth, humid gray earth,earth with vegetation, humid gray land. The paperis structured in: an introduction of the importanceon self-organizing maps and adaptative resonancenetwork for the classification of images; andescription of the algorithms used by the two neuralnetworks; information of the problem; the use of theneural networks SOM and ART2 in the classificationof satellite images; and obtain to conclusions andfuture works.