Color based K-Means Clustering For Image Segmentation to Identify the Infected Leaves

Abstract

One of the major process in image processing is Segmentation. Every image is framed with the collection features. The image is split into small components called segments, based on the common features. In this proposed work, the infected leaves are considered as input. Image analysing steps such as, image pre processing is done for Image enhancement, image reshaping is done for converting the two-dimensional colored data, represented as R(m×n) G(m×n) B(m×n)data into one-dimensional R(mn×l)G(mn×l)B(mn×l) data. This will generate a matrix of size mn×3, the three columns are defined to represent the RGB data as ri,giand bi, i=1,2,…mn. RGB features are used for segmenting the leaf image by using K-Means clustering algorithm, with various distance measuring functions. The clustering is done on the pixel color. Totally, three colors, Green, Yellow or white (light colors) and Red or brown color are clustered as three segments. Finally, the given image is split into three sub images (segments), based on the colors, which will help to identify the infected leaves based on the colors in agricultural field.