Multisource Classification of Remotely Sensed Data Based on Bayesian Data Fusion Method

In this paper, a new method for classification of multisource data is proposed. The images formation model, contextual model and reliability factors are taken into account in the method. The performance of the method is evaluated by fusing Landsat TM images and SAR image for land-use classification. Significant improvements in classification accuracy compared to the SAR image classifier are obtained. So it is an effective and robust method for multisource classification of remotly sensed data.