POLARIMETRIC CONTRAST CLASSIFICATION OF AGRICULTURAL FIELDS USING MAESTRO-1 AIRSAR DATA

In this paper, we show the results of a polarimetric contrast classifi er applied to the MAESTRO 1 JPL-AIRSAR data of agricultural fields. Fo r the Dutch Flevoland agricultural site, we first compare the classifi er's performance to results given in earlier studies, where polarimetr ic information was not available. This comparison shows that the avail ability of the measurements of the full polarimetric covariance matrix greatly enhances single-date classification accuracy. In the next ste p, we illustrate the use of frequency diverse classification results f or further separation within specific classes. Furthermore, we will sh ow the use of the classifier in two unsupervised approaches. In the fi rst approach, we define the training clases from theoretical and semi- empirical models and run the classifier on a fully calibrated Flevolan d dataset. In the second approach, we take covariance matrices extract ed from the Flevoland data as the training set and use these in an uns upervised classification of the U.K. Feltwell agricultural site.