Arecanut Disease Classification using CNN

Abstract

Arecanut is a tropical crop, which is popularly known as betel nut. India ranks second in producing and consuming arecanut in the world. Throughout its life cycle, it is affected by a variety of diseases, from root to fruit. The current approach for detecting diseases is simply observation with the naked eye and farmers have to carefully analyse each and every crop periodically to detect the diseases. In this paper, a new system is proposed which helps in detecting the diseases of arecanut, leaves, and its trunk using Convolutional Neural Networks and suggests remedies for it. To train and test the CNN model, Dataset is created which consists of 200 images of arecanut both healthy and diseased. The train and test data are divided into a ratio of 70:30. For compilation of model categorical cross-entropy is used as loss function with adam as optimizer function and accuracy as metrics. A total of 50 Epochs are used to train the model to achieve high validation and test accuracy with minimum loss. The proposed approach was found to be effective and 81.35 percent accurate in identifying the arecanut disease.