An Integrated Optimization Model for Plant Diseases Prediction with Machine Learning Model

Abstract

The endurance of people is for the most part founded on the legitimate efficiency of farming. The paddy plant is considered as a significant establishing crop in working on the conservative level of our country. These days, the yield level of paddy harvest may be limited because of a few sicknesses. The means engaged with the picture handling are picture securing, pre-handling, and division of picture, highlight extraction and arrangement. This paper considered the four unique sorts of rice plant illnesses in particular, bacterial curse, sheath decay, earthy coloured spot and impact sicknesses are recognized and ordered in this proposition with the assistance of picture handling and AI procedures. The developed model comprises of the K-implies grouping is utilized for division of the sick and solid part of the leaves. The highlights of variety and surface are extricated in the phase of element extraction. Four distinct classifiers specifically, ANN, DNN, KNN and advanced DNN classifier with Jaya enhancement are utilized in this proposition for distinguishing and ordering the event of illnesses in paddy goes out climate. The performance analysis of the detection technique in the greenhouse environment presents better outcomes than the rice crops which are cultivated in the agricultural environment. In addition to this, ANN and KNN classifiers are applied for detecting the blast disease of the rice crop; whereas, optimized DNN classifier with Jaya optimization algorithm detects and few more classifies diseases. The output performance metrics such as TPR, TNR, FPR, FNR, NPV, FDR and F1-score, accuracy, precision, and recall are determined for each classifier and compared with the existing algorithms