An Ensemble Methods based Machine Learning Approach for Rice Plant disease diagnosing

Abstract

Even though the annual rice production decreases by 37% because of rice plant diseases still there isn’t any proper application developed which can identify rice plant diseases accurately and instruct farmers to control the spreading of rice plant diseases. This might be solved by creating a software program for farmers that can diagnose rice plant illnesses and provide instructions to farmers to do treatments for those ailments. Convolutional Neural Networks, Which are particularly efficient in picture recognition and classification, may be employed directly for this illness detection procedure. In this study, the author has created an ensemble model Which can identify rice plant diseases accurately. This has been created by integrating VGG-16, Alex Net, and ResNet_50 models Which ha identified in the study[1]. Under this study, the author has evaluated the accuracies of all these three modules individually and he could able to get 98.50 %, 94.33 %, and 99.84 % for AlexNet, VGG_16, and ResNet_50 respectively. In this ensemble model, it considers confidence as a parameter and uses it to measure the accuracy of the predicted results of disease-affected rice plant leaves. This model will help farmers to identify rice plant diseases effectively.