Identification of Disease in Cassava Leaf using Deep Learning

Abstract

Automating the disease detection in plants is one of the most complex recent challenges faced by agricultural experts and farmers worldwide. The traditional laboratory testing methods are inefficient for detecting diseases in crops such as cassava. Unlike rice and maize, cassava is the third-largest source of carbohydrates. It is nutritious, it consists of resistant starch and its root is high in vitamin C. These plants suffer from four major diseases which spread to neighboring cassava plants and affect the cultivation. This paper describes the work done to detect and classify the disease, which will help in figuring out if the crop is healthy and can prevent further spread of disease. Computer vision is a subset of deep learning, which trains computers to interpret and understand the visual world. The paper discusses various ways for training models and their results for disease classification. The work achieves the best accuracy of 89.01% by using the EfficientNetB3 model.