Unsupervised Classification of Remote Sensing Data Using Graph Cut-Based Initialization

In this paper we propose a multistage unsupervised classifier which uses graph-cut to produce initial segments which are made up of pixels with similar spectral properties, subsequently labelled by a fuzzy c-means clustering algorithm into a known number of classes. These initial segmentation results are used as a seed to the expectation maximization (EM) algorithm. Final classification map is produced by using the maximum likelihood. (ML) classifier, performance of which is quite good as compared to other unsupervised classification techniques. © Springer-Verlag Berlin Heidelberg 2005.