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.
KeywordsExpectation Maximization Algorithm Multi Spectral Image Remote Sensing Data Remote Sensing Image Ground Truth Information
- 3.Gath, I., Geva, A.B.: Unsupervised Optimal Fuzzy Clustering. IEEE PAMI 11, 773–780 (1989)Google Scholar
- 4.Haykins, S.: Neural Networks. Pearson Education, London (1999)Google Scholar
- 5.Kilpatrick, D., Williams, R.: Unsupervised Classification of Antarctic Satellite Imagery Using Kohonen’s Self-Organising Feature Map. In: Proceedings of IEEE International Conference on Neural Networks, vol. 1, pp. 32–36 (1995)Google Scholar
- 6.Moon, T.K.: The Expectation-Maximization Algorithm. IEEE Signal Processing Magazine, 47–60 (November 1996)Google Scholar
- 8.Felzenszwalb, P.F., Huttenlocher, D.P.: Efficient Graph-Based Image Segmentation. International Journal of Computer Vision 59(2) (September 2004)Google Scholar