Classification of Wheat Crops from ASAR Data

In this study, the use of ASAR HH/VV backscatter ratio for winter wheat classification has been assessed. The experimental analysis has been conducted over an agricultural site located in Southern Italy and mostly dedicated to durum-wheat cultivation. Two temporal series of ASAR AP data acquired during the 2006 and 2007 growing seasons have been analyzed. The obtained results show that a wheat/non-wheat classification accuracy of 75-80% can be achieved by using a single ASAR AP image, acquired during the peak of the wheat growing season. Higher classification accuracies, approximately equal to 90%, can be obtained when a priori information on field boundaries is available and speckle filtering is carried out at field scale.