Green Tech: An Android Application for the Automatic Identification of Potato Leaf Diseases using Deep Learning.

Abstract

Abstract: Research on sustainable agricultural development is becoming more and more significant owing to the advancements in agricultural technology and the use of artificial intelligence to diagnose plant diseases. Potato is an extensively consumed food corp worldwide, with India being one of the biggest producers. However, numerous diseases, particularly leaf diseases like early blight and late blight, have a substantial negative influence on the quality and quantity of potatoes. Manual interpretation of these diseases is inconvenient as the process is time-consuming and requires high expertise. Thus, the researchers are working on automatizing leaf disease detection in real time. In this work, we have presented an Android application for classifying a potato leaf into - healthy, early blight or late blight categories. A deep learning model based on a pre-trained VGG 16 model was designed for classification purposes. Transfer learning is also applied in detecting leaf diseases in several other plants. The model was fine-tuned using a dataset containing 1500 potato leaf images with 900, 300 and 300 train, test and validation images respectivelyFinally, the model was converted into a TFlite file for integration and deployment in Android Studio. Experiments showed promising results with an obtained accuracy of 98%.