An investigation on deep convolutional Neural network models for tomato Leaf disease identification

Abstract

Plant diseases are the most catastrophic factor in the agriculture newlinesector and cause a significant reduction in yield and economic loss. Tomato is newlinethe most commonly cultivated vegetable crop worldwide due to its rich newlinenutrition and various health benefits. Many experts consider a disease to be a newlinesignificant hazard to tomato cultivation. The majority of the crops are highly newlineaffected by multiple diseases, which causes tremendous loss to the farmers newlineand the agricultural economy. Thus, accurate detection of these diseases is newlinehighly preferred in the agriculture field. It becomes challenging to control the newlinespread of disease over crops and ensure the minimization of production loss. newlineIn traditional methods, human experts in the agricultural sector newlinehave been indulged in finding out the anomalies in tomato plants caused by newlinediseases. Moreover, the conventional techniques consume more time, and it is newlinea complex task. Similarly, farmers face a challenging burden in keeping track newlineof their plants to avoid the spread of disease. Therefore, a system that newlineperforms automatic, rapid, and precise leaf disease detection is necessary to newlineidentify infections early and be of great significance. A computer newlinevision-based system is developed to enable a machine learning approach for newlineautomatic identification of plant disease in the agriculture field. Machine newlinelearning, known as Deep Learning, focuses on developing systems that newlineautomatically extract features from raw data. newlineConvolution Neural Network (CNN) is one of the most popular newlinemethods of deep learning. Recently, CNN models are most widely used in newlineseveral agricultural problems like plant/crop disease recognition, fruit newlineclassification, weed detection, and pest identification. newline