A MRF based segmentatiom approach to classification using Dempster Shafer fusion for multisensor imagery

A technique has been suggested for multisensor data fusion to obtain landcover classification. It takes care of feature level fusion with Dempster-Shafer rule and data level fusion with Markov Random Field model based approach vis-a-vis for determining the optimal segmentation. Subsequently, segments are validated and classification accuracy for the test data is evaluated. Two illustrations of data fusion of optical images and a Synthetic Aperture Radar (SAR) image is presented and accuracy results are compared with those of some recent techniques in literature for the same image data.