Leaf Disease Detection Using Trained Convolutional Neural Network

Authors

Jameer kotwal, Dr.Ramgopal Kashyap
Amity University Chhattisgarh, India.

Dr.Pathan Mohd Shah A Karim
Smt.Kashibai Navale Engg. College.

Abstract

India is the largest agricultural exporter worldwide. Plant protection from various diseases is an essential factor. According to the survey, 10 -12% of crop production gets the loss. Identifying the plant disease is the argument of the prevention of the plant disease efficiently and effectively in various environments. Tomato is one of the most vegetables worldwide. As it a concise duration spam crop and produces high demand, it makes it economically better, and the area under cultivation is increasing daily. Still, some tomato diseases can cause damage to this crop. Early blight, Bacterial leaf spot, tomato mosaic virus, and leaf curl are common tomato leaf diseases that seriously affect tomato yield. To act against the problem, early detection of diseases using a deep learning technique will benefit the farmer. Here the tomato disease dataset contains( 9 conditions & one healthy leaf of tomato) taken from the plant Village dataset. The leaf images trained & tested deep learning architecture, namely VGG16, VGG19, InceptionV3, and ResNet50. Based on batch size, learning rate and trained weights the accuracy and loss are analyzed on a different architecture.