Vegetable and Fruit Leaf Diseases Detection using ResNet

Abstract

Agriculture is one of the main factors that decides the growth of any country. India is an agricultural country which has the majority of the population dependent on agriculture as their main income source. Having diseases is quite natural in plants due to changing climatic conditions and environmental conditions. Diseases obstruct the growth of plants and affect their production. Due to this, it is very important to detect the diseases as it may infect other plants. In this paper, we aim to propose a solution to detect the diseases from various vegetables and fruits from the PlantVillage dataset using the ResNet algorithm. Out of the total 38 classes, plant leaves are classified into 26 diseased classes or 14 healthy classes. As a result, test accuracy obtained is 99.2%.