A Fuzzy Approach of Large Size Remote Sensing Image Clustering

Remote sensing image segmentation is a very important stage in remote sensing image processing. In many different segmentation techniques such as KMeans, C-Means, Watersed .... KMeans is one of the widely used algorithms for remote sensing image segmentation. However, This algorithm considers only pixels that belong to the nearest cluster. The Fuzzy C-Means algorithm fixed the problem of KMeans algorithm by considering that a pixel can belong to multiple clusters with corresponding dependency. However, the Fuzzy C-Means algorithm has a very slow execution speed. In particular, the Fuzzy C-Means algorithm has memory problems when doing clustering of large images such as remote sensing images. This article presents the new remote sensing image clustering algorithm MapReduce Fuzzy to overcome the disadvantages of Fuzzy C-Means’ computation time and memory problem when executing on large image without reducing cluster quality.