Analysis of Automatic Classification and Diseases Identification of Plant Diseases using Pre Train Convolutional Neural Network Models

Every country’s economy depends upon agriculture growth, but every year there is a huge loss in production due to plant diseases, so there should be a system for early identification of plant diseases to prevent losses. As there is good availability of smartphone with farmers now various deep learning models based applications can automatic classify the plant diseases. In this paper of various deep learning pretrained CNN models have been analyzed on 15 classes of potato, tomato and pepper plants from Plant Village dataset.The validation accuracy of VGG16 produced best results as 98.10 where,Xception, InceptionV3 andInception-ResNet v2 ,NASENetMobile also shown good accuracy on validation data. Various other factors related with CNN models like performance, testing on real time data and future scope also elucidated.