Integration of multi-temporal SAR and optical images for the classification of agricultural crops

Food is one of the most basic factors of our physical and intellectual wellness and is fundamental to drive forward the economic and social progress of people. Spatial and temporal variability of food availability creates economic, social and political exchanged and facts tensions between different countries. Remote sensing offers an efficient and reliable means of information in order to map crop types, to understand their growth evolution and to provide structure information about the health of the vegetation. In this context, the objective of this thesis work is to assess the improvements of in the crop classification accuracy by exploiting a combination of optical and SAR data, evaluated on by considering data sets acquired over extended agricultural crop fields during the AgriSAR 2009 campaign in Flevoland. The optical images from visible to infrared wavelengths offer different information on vegetation properties.SAR can provide complementary information to the optical data. Combining the information from these two types of sensors increases the available information for distinguishing each target class and its respective signature.

Integrazione di immagini multi-temporali SAR e ottici per la classificazione di colture agricole. Food is one of the most basic factors of our physical and intellectual wellness and is fundamental to drive forward the economic and social progress of people. Spatial and temporal variability of food availability creates economic, social and political exchanged and facts tensions between different countries. Remote sensing offers an efficient and reliable means of information in order to map crop types, to understand their growth evolution and to provide structure information about the health of the vegetation. In this context, the objective of this thesis work is to assess the improvements of in the crop classification accuracy by exploiting a combination of optical and SAR data, evaluated on by considering data sets acquired over extended agricultural crop fields during the AgriSAR 2009 campaign in Flevoland. The optical images from visible to infrared wavelengths offer different information on vegetation properties.SAR can provide complementary information to the optical data. Combining the information from these two types of sensors increases the available information for distinguishing each target class and its respective signature.