Farmer Friendly Smart App for Pomegranate Disease Identification

Abstract

Pomegranate is a fruit with a good yield that grows in several Asian countries and is the most profitable one. However, due to a variety of factors, the plants become affected by a wide range of illnesses, resulting in the full destruction of the plant and a drastically reduced yield. Preventing decreases in agricultural production is possible with the early detection of plant diseases. Pomegranate leaf diseases are extremely tough to keep track on manually. As a result, pomegranate plant diseases are detected using Deep Learning (DL). Automating the disease detection system for pomegranates using leaf images is the goal of this study. Image gathering, processing, classification, and deployment are all part of the disease detection system process. Pomegranate leaf health and disease images are built using Mendeley data. The raw image is then processed further. Two DL models, AlexNet and VGG-16, are employed for classification. Accuracy and loss metrics are used to identify the optimal model. The metrics analysis shows that AlexNet is efficient in detecting leaf disease. A mobile app utilizing the AlexNet approach is then created to assist farmers in the detection of pomegranate disease without the assistance of specialists.