Segmentation of satellite image by enhanced spatial clustering approach

Image Segmentation is an effective technique which partition an image as essential regions or objects and the segmented regions are further used for computer vision process. This project provides an image segmentation approach to accurately detect boundary regions of agriculture fields in satellite image. Here image is first enhanced using contrast limited adaptive histogram equalization which divides the contextual regions and the histogram of an image equalized in every region. Spatial information of the image is obtained by edge detection method in order to neglect noises in the image while clustering. Then spatial information supplied to fuzzy c-means algorithm to cluster the image. The proposed approach uses spatial information to segment the satellite image which makes segmentation efficient. Thus, the experimental result shows that, the proposed spatial clustering scheme, segment the images with efficient boundary detection and neglects incorrect clusters in the image.