A Mobile Application for Plant Disease Classification using Convolutional Neural Network

Mobile applications have now become the most efficient and direct way to deliver information to the user. Agriculture requires mobile applications that can identify plant diseases from pictures of a plant’s leaves. A model with fewer parameters is taken into consideration while implementing this from the existing models. A mobile application built on the Mobilenetv2 architecture is created to categorise plant diseases from leaf photos. The plant village dataset, a publicly accessible dataset that includes photos of both healthy and diseased plant leaves from crops grown all over the world, is taken into consideration for training a model. There are 54,305 samples of fourteen crops over 38 groups in the collection. Of these, 10,849 samples are considered for testing and the remaining 43,456 samples are used for training and validation. The proposed model has a 92.45% accuracy rate. The model has been implemented as a mobile application that will detect plant disease. This application will be an easy way for the farmers to detect the diseases and get a remedy at the earliest.