Utilizzo di Reti Neurali, Algoritmi genetici e SVM applicati ad immagini telerilevate per la classificazione di discariche incontrollate

In the field of the environmental protection, uncontrolled landfills are a hazard to the health and safety of residents strongly threatening the territory because of the heavy impact on the landscape and environment.
As is known, the use of remote sensing can make a valuable contribution to the monitoring of the territory through the implementation of dedicated algorithms for the identification and classification of such abuse.
In this perspective, this article intends to contribute through comparison of different methodologies (support vector machines / processes of self-organizing neural networks SOM / and genetic algorithms) to search and classify potential landfill sites in Calabria.
Specifically, we paid particular attention to the genetic algorithm used to solve difficult problems with the objective function that do not satisfy the Lipschitz condition. These algorithms, in searching of the best solution, manipulate a family or a population of solutions by implementing the strategy of survival of the fittest.