Fusion of support vector machines for classifying SAR and multispectral imagery from agricultural areas

A concept for classifying multisensor data sets, consisting of multispectral and SAR imagery is introduced. Each data source is separately classified by a support vector machine (SVM). In a decision fusion the outputs of the preliminary SVMs are used to determine the final class memberships. This fusion is performed by another SVM as well as two common voting schemes. The results are compared with well-known parametric and nonparametric classifier methods. The proposed SVM-based fusion approach outperforms all other concepts and significantly improves the results of a single SVM that is trained on the whole multisensor data set.