Please use this identifier to cite or link to this item:
|Title:||On the systematic implementation of artificial neural networks in the classification of variance images and shoreline extraction|
|Authors:||RIGOS Anastasios; VOUSDOUKAS MICHAIL; ANDREADIS Olympos; TSEKOURAS Georgios; VELEGRAKIS Adonis|
|Citation:||FRESENIUS ENVIRONMENTAL BULLETIN vol. 23 no. 11 p. 2677-2686|
|Publisher:||PARLAR SCIENTIFIC PUBLICATIONS (P S P)|
|Type:||Articles in periodicals and books|
|Abstract:||Monitoring of the shoreline position has become an issue of urgency given the high socio-economic impact due to the population density of the coastal zone, the increasing erosion and the projected sea-level rise. In this study, we implement a specialized monitoring system to generate a database consisted of variance images also called SIGMA images. We, then, apply a radial basis function (RBF) neural network trained with the aid of fuzzy clustering. The RBF network is able to elaborate on the image histograms in order to perform efficient image classification and shoreline extraction. The outcoming contributions of the current study can be summarised as follows. We develop a specialized regression strategy to approximate the image histograms, where for each histogram we extract a set of regression parameters. Then, the key idea is to use these parameters as the input data of the RBF network and associate them with a set of intensity thresholds that directly quantify the corresponding shorelines. In addition, the utilization of fuzzy cluster analysis provides the means to effectively estimate all the network’s parameters by taking into consideration any uncertainty hidden in the histograms. The simulation experiments showed that the proposed algorithmic framework exhibits an accurate behaviour, while it also constitutes a fully automated process.|
|JRC Directorate:||Sustainable Resources|
Files in This Item:
There are no files associated with this item.
Items in repository are protected by copyright, with all rights reserved, unless otherwise indicated.