COLOUR BASED IMAGE SEGMENTATION USING FUZZY C-MEANS CLUSTERING

Abstract

Mostly due to progresses in spatial resolution of satellite imagery, methods of segment- based image analysis for generating and updating geographical information are becoming more and more important. This work presents a new image segmentation based on colour features with Fuzzy c-means clustering unsupervised algorithm. The entire work is divided into two stages. First enhancement of color separation of satellite image using decorrelation stretching is carried out and then regions are grouped into a set of five classes using Fuzzy c-means clustering algorithm. Using this two step process, it is possible to reduce computational cost avoiding feature calculation for every pixel in image. Although colour is not frequently used for image segmentation, it gives a high discriminative power of regions present in image. In remote sensing, process of image segmentation is defined as: the search for homogenous regions in an image and later classification of these regions. It also means partitioning of an image into meaningful regions based on homogeneity or heterogeneity criteria. Image segmentation techniques can be differentiated into following basic concepts: pixel oriented, Contour-oriented, region-oriented, model oriented, color oriented and hybrid. Color segmentation of image is a crucial operation in image analysis and in many computer vision, image interpretation, and pattern recognition system, with applications in scientific and industrial field(s) such as medicine, Remote Sensing, Microscopy, content based image and video retrieval, document analysis, industrial automation and quality control. The performance of color segmentation may significantly affect quality of an image understanding system .The most common features used in image segmentation include texture, shape, grey level intensity, and color. The constitution of right data space is a common problem in connection with segmentation/classification. In order to construct realistic classifiers, features that are sufficiently representative of physical process must be searched. In literature, it is observed that different transforms are used to extract desired information from remote-sensing images or biomedical images. Segmentation evaluation techniques can be generally divided into two categories (supervised and unsupervised). The first category is not applicable to remote sensing because an optimum segmentation (ground truth segmentation) is difficult to obtain. Moreover, available segmentation evaluation techniques have not been thoroughly tested for remotely sensed data. Therefore, for comparison purposes, it is possible to proceed with classification process and then indirectly assess segmentation process through produced classification accuracies. For image segment based classification, images that need to be classified are segmented into many homogeneous areas with similar spectrum information firstly, and image segments' features are extracted based on specific requirements of ground features classification. The color