AN IMAGE PROCESSING AND NEURAL NETWORK BASED APPROACH FOR DETECTION AND CLASSIFICATION OF PLANT LEAF DISEASES

In the present paper we propose and evaluate a framework for detection and classification of plant leaf/stem diseases using image processing and neural network technique. The images of plant leaves affected by four types of diseases namely early blight, late blight, powdery-mildew and septoria has been considered for study and evaluation of feasibility of the proposed method. The color transformation structures were obtained by converting images from RGB to HSI color space. The K- means clustering algorithm was used to divide images into clusters for demarcation of infected area of the leaves. After clustering, the set of color and texture features viz. moment, mean , variance, contrast, correlation and entropy were extracted based on Color Co-occurrence Method (CCM). A feed forward back propagation neural network was configured and trained using extracted set of features and subsequently utilized for detection of leaf diseases.