Disease Detection in Paddy Crops of Rural India Using Machine Learning Techniques

Abstract

newlineThe survival of human beings is generally based on the proper productivity of agriculture.The paddy plant is considered as a major planting crop in improving the economical level of our country. Nowadays, the yield level of paddy crop might be minimized due to several diseases. Bacteria, fungi, virus and certain harmful insects are the main causative agents for such disease occurrence on the paddy crop. The diseases which affect the early stage of the paddy crops influences in the whole stage of crop cultivation. In early days of agriculture, the manual detection of diseases have been carried out by farmers. Naked eye inspection is commonly performed in manually detecting the diseases in the crops. In this primitive method, enormous time is needed for classifying the diseases and normally leads to certain errors. Image processing is one of the emerging techniques for identifying and classifying the different types of diseases and it overcomes the issues encountered during the manual detection of diseases. Image processing technique solves several issues involved in the cultivation of crops including, recognition and classification of plant diseases, discrimination of certain weeds and disease forecasting. newlineThe steps involved in the image processing are image acquisition, pre-processing, and segmentation of image, feature extraction and classification. Four different types of rice plant diseases namely, bacterial blight, sheath rot, brown spot and blast diseases are identified and classified in this thesis with the help of image processing and machine learning techniques. K-means clustering is used for segmentation of the diseased and healthy portion of the leaves. The features of colour and texture are extracted in the stage of feature extraction. Four different classifiers namely, ANN, DNN, KNN and optimized DNN classifier with Jaya optimization are used in this thesis for identifying and classifying the occurrence of diseases in paddy leaves in both agricultural and green house environment.