Analytical Method of Multi-Objective Genetic Algorithm with Multi-Objective Messy Genetic Algorithm in Satellite Image Segmentation

Abstract

Image can be dividing into different Segmentation. In image processing , the important task is Segmentation process methods. This method involves such as K-means clustering, watershed segmentation, Fuzzy c-Means, Iterative Self Organizing Data. Clustering methods depends powerfully on the selection of the primary spectral signatures which represents initial cluster centers. Normally, this is either done physically or erratically based on statistical operations. In this case the outcome is random and sometime inaccurate. In base paper an unsupervised method based on Multi-Objective Genetic Algorithm (MO-GA) for the selection of spectral signature from satellite images is implemented. The goal is to make greatest cluster centers as an initial population for any segmentation technique. Experimental results are conducted using high-resolution SPOT V satellite image and the verification of the segmentation results is based on a very elevated resolution satellite image of kind Quickbird. The spectral signatures method to Fuzzy c-means and K-means by MO-GA method increased the speed of the clustering algorithm to approximately4 times the speed of the random based selection of signatures. In this paper unsupervised method is comparative with Multi-Objective Messy Genetic Algorithm(MOMGA) with existing MO-GA methods for the selection of spectral signature using satellite images.