Sugarcane Plant Disease Detection Using Transfer Learning

Abstract

Sugarcane farming is the well-organized type of farming. Sugar and ethanol are produced mostly from sugarcane. We know, Bangladesh is an agricultural country, and agriculture is the backbone of the Bangladeshi economy. Plant diseases are the cause of significant economic losses in the agriculture business around the world and for Bangladesh purpose sugarcane plant disease is one of them. It is one of our farmers' most well-known issues. This study involves numerous plant diseases in order to discover a remedy and better strategies to reduce them. The dataset consists of images of healthy as well as diseases of sugarcane plants. For this research nearly 2000 images have been collected. In this study, transfer learning based popular image classifiers, such as, ResNet50, VGG19, and Incep-tionV3 were considered which aid in the classification and detection of illness images. Among those classifiers, ResNet50 outperformed the other two algorithms with an accu-racy of 99.75% on test set.