Performance Analysis of Cotton Leaf Disease Detection System

Abstract

In India, agriculture is the primary source of farmers' revenue. India's most widely grown and traded crop is cotton. It allows farmers to make good capital and will boost their revenue. Cotton's vulnerability to many diseases is a significant issue. Plant illnesses must be detected as soon as feasible to prevent productivity loss. A method for automatic disease detection will be needed for this. In this study, we suggest an automated process that uses deep learning techniques to identify prevalent illnesses that affect cotton leaves. One of Ethiopia's most significant crops in terms of economic importance is cotton; however, there are numerous limits to its use in some areas. Typically, these are limited to detecting the majority of leaf diseases. In this project, we used fungus databases for training. For prediction, we propose using the modified CNN to classify different types of fungus. The fungus dataset consists of 1951 images divided into four classes.