IoT based Wheat Leaf Disease Classification using Hybridization of Optimized Deep Neural Network and Grey Wolf Optimization Algorithm

Abstract

Image classification has become a hot research area in identifying diseases in plants. For eliminating the financial loss of farmers, plant diseases identification by image processing models can be used to save the agricultural products. This paper presents a new IoT and cloud based classification model of wheat leaf diseases using optimal deep neural network (ODNN) model. The proposed model comprises of four processes, viz., i) acquisition, ii) preprocessing, iii) segmentation, iv) feature extraction and classification. Initially, acquisition of images takes place using Internet of Things (IoT) devices. Once the image is preprocessed, application of K-means clustering is made that extracts the diseased areas in the leaf images. Next certain unnecessary green areas are discarded from diseased area utilizing thresholding technique. Then, a set of features such as color of the leaf, texture of the leaf and shape of the leaf are extracted in the feature extracting process. At the end, ODNN model has been applied for proper image classification, which employs the Grey Wolf Optimization (GWO) Algorithm for the parameter optimization of DNN. A series of experimentation is conducted to verify the effectiveness of the presented DNN-GWO model on the identification of wheat leaf diseases.