Potato Leaf Disease Detection with Deep Learning Method Using Leaf Images and CNN Approach

Abstract

Potato leaf disease is one kind of venom disease which is the reasons of young leaves to roll and turn yellow or pink. Early blight, late blight, septoria blight etc are the familiar types of potato leaf diseases which spectacle the syndrome on the affected zone and spoil the whole potato gradually. Appearance of potato leaf diseases depends on weather. These kinds of disorder syndrome can be cause of qualitiless production and even can slow down the farmers economy. By detecting this indisposition on primary stage can bring off farmars from this suffering. Identifying this disorder by experts is not always affordable on lab specially in rural area. We promote a model that identify automatically affected leaf conscientiously in a short time . In this paper, a combined (CNN-LSTM) model along with a traditional machine learning algorithms employing MR images have proposed to identify disease on potato leaf. we have used dataset of 12,009 potato leaf images, 7,376 images of them are affected leaf. Heathy & disorder-impacted leaf both were used to provide a separate layer between abnormal and normal aspects.The proposed CNN-LSTM model, is succeed in (keras and tensorFlow) and its ability is surpassing . This Convolutional neural networks(CNN) model can detect affected potato leaf in short time with the accuracy of 97.78% and it’s quite reliable than other models.