The role of semi-automatic classification techniques for mapping landscape components. The case study of Tratturo Magno in Molise region

The purpose of this research is to test different mapping techniques that use semi-automatic recognition methods of the contents of multispectral satellite imagery, to understand how they may support the work of analysis and interpretation of the landscape crossed by the Tratturo Magno between the internal areas of Molise and the Adriatic coast. Three different approaches have been tested with GIS and Google Earth Engine; the first one concerns the computing of Vegetation Indices; the second one is about the unsupervised classification method and the third one is about the supervised classification method. The methodology has been applied, on one hand to one single satellite image through the Semi-automatic classification Plugin in GIS and on the other hand to a series of satellite imagery spanned over a period of time through Google Earth Engine. However, no one of the presented methods is not fully matching the correct procedure to uniquely map the features of the landscape crossed by the tratturo.