Wheat Crop Mapping by using ASAR AP data at High Incidence Angles: Foggia case study

The objective of this paper is to investigate the use of HH/VV backscatter ratio, derived from data acquired by the Advanced Synthetic Aperture Radar (ASAR) system at high incidence angles (i.e. swaths I5-I7), to discriminate wheat from other crops. The wheat classification is obtained by applying an optimal threshold derived from the statistical distribution of the co-polarized backscatter ratio. The classification is firstly applied to fields monitored during the experimental campaign, then it is extended to the entire study area and compared with a classification obtained from multi-temporal SPOT images. The experimental data set consists of a temporal series of 7 ASAR AP and 3 SPOT images acquired in 2006 over an agricultural site located in the Capitanata plain, close to the Foggia town (Southern Italy). The analysis shows that it is possible to obtain classification accuracies in the wheat and non-wheat discrimination higher than 85% under two conditions: 1) ASAR AP data must be acquired during the heading phenological stage (e.g. from the end of April to end of May); 2) the HH and VV backscatter need to be estimated at field scale (corresponding to an estimated number of looks higher than 65).