An individual grape leaf disease identification using leaf skeletons and KNN classification

The most challenging process in agricultural applications is identification of leaf individually. In this paper, the classification of grape leaf diseases is proposed along with the leaf identification. Initially, the leaf skeletons are identified based on grape images. Since, the leaf skeletons are used for estimating the positions and directions of the leaves. The Tangential Direction (TD) based segmentation algorithm is proposed for retrieval of skeletons. If the grape leaf images are classified, then the histograms of H and a color channels are generated and the pixels values are observed to distinguish the healthy and diseased tissues. Then, extract the features and classify by using the KNN classification algorithm in order to find the leaf diseases.