Segmentation of Defected Regions in Leaves using K-Means and OTSU's Method

Image segmentation is the process of dividing a digital image into multiple fragments. The main theme of segmentation is to represent or convert the image to simplest form for further processing. Image segmentation is most widely used to locate objects and boundaries in images. More clearly, image segmentation is the process of assigning an address to every pixel in an image, so that the pixels having same property can share certain characteristics. In existing method, thresholding and histogram techniques are used for segment the defected regions in leaves. Thresholding is one of the segmentation method which is used for segmenting the image by fixing a threshold value and histogram is a method used for collecting the information about the image from its background. The drawback of this method is that it does not provide proper segmentation region. So in order to overcome these drawbacks in the proposed, K-means algorithm and Otsu's methods are used to segment the defected regions in leaves. In K-means segmentation algorithm, defected image is grouped into eight clusters which depends on its intensity and an iteration process. Each and every time the cluster center value is changed depends upon the minimum Centroid value. Finally the given image is segmented in order to identify its defected regions. The next is the Otsu's method is one which is used to segment the image by automatic thresholding process. Each and every time in the process, the threshold value is changed based on its mean and variance value, at last the maximum variance value is considered as the threshold value. Then the proposed method provides better segmentation result when compared with the existing methods in terms of defect identification with accuracy.