Publication

Abstract : Identification of crop diseases is a challenging task in the agricultural field. Various methods are proposed for classification and disease identification. Considering all the researches done so far, accuracy is the major problem identified. It is due to the lack of clarity in obtaining the image of the crop from different viewpoints. A preprocessing method is implemented for the captured image and is imperiled to an improved clustering algorithm, specifically k-means algorithm, to acquire the disease-ridden part of the leaf. The classification of the diseases in the leaf is done by using Bayesian-based SVM classifier. The identified diseased portion might be given on the way to a processing method in order to expand the most affected area. Some part like the plague-ridden part of the leaf might get exposed toward a combined histogram algorithm called histogram of oriented gradient (HOG) algorithm and region of interest (ROI) algorithm to segregate the features. For obtaining the extracted features and to classify and identify the plant diseases, the proposed unsupervised machine learning method (SVM) is better than previous techniques.