Deep Learning for Detection of Mango Leaf Disease: A Comparative Study Using Convolutional Neural Networks Models

Abstract

The diagnosis of diseases in mango leaves presents a complex challenge, compounded by the diversity of crop types, variability in agricultural disease indicators, and a multitude of environmental factors. Early detection of these diseases is particularly difficult, as existing methods often rely on data limited to specific geographic areas, thus restricting their efficacy. The timely identification and management of such diseases are crucial in averting substantial financial losses for farmers. This research introduces an innovative approach, leveraging image processing technology to detect diseases in mango leaves. The study employs a Convolutional Neural Network (CNN) model, specifically the Proposed model, which has shown remarkable accuracy of 97.92% in this context. This model was rigorously tested on both normal and diseased mango leaves, showing its efficacy in differentiating between healthy and unhealthy foliage. The application of this algorithm to leaf images facilitates the categorization of mango tree leaves into healthy or diseased categories, offering a timely and efficient solution for early disease detection. This novel approach not only promises to enhance disease management in mango cultivation but also sets a precedent for applying similar techniques in broader agricultural contexts, potentially revolutionizing plant disease diagnosis and prevention.