Satellite Image Processing Using Fuzzy Logic and Modified K-Means Clustering Algorithm for Image Segmentation

Abstract

Satellite images are useful in providing a real time dynamic picture of the earth and its environment. The large assemblage of remote sensing satellites orbiting the earth provide an extensive and periodic coverage of the planet through the capture of live images round the clock, in turn enabling numerous uses for the benefit of mankind. In the field of satellite image processing, image segmentation is one of the vital steps for extracting and gathering huge amount of information from the satellite images. The basic k-means clustering algorithm is simple and fast in terms of dealing with the required segmentation, but the limitation associated with this clustering is its inability to produce the same result for every run, as the resulting clusters depends on the initial random assignments. In this paper, an enhanced modified k-means clustering algorithm is proposed for the effective segmentation of the satellite images with an objective to overcome the demerits of the traditional k-means by combining fuzzy logic with the membership function. The proposed methodology continuously produces the same result for each run. As an outcome, the experimental results proved that the enhanced k-means algorithm is an effective and more efficient process for the precise and accurate segmentation of satellite images. Index Terms : Image Segmentation, Satellite Imagery, Fuzzy logic, K-Means, Clustering.