Detecting Mango Leaf Disease with VGG 16 Transfer Learning Model

Abstract

Mango leaf disease is a major issue affecting worldwide mango trees. If left untreated, it can cause substantial damage to the leaves and potentially cause the tree to die. The importance of early discovery in controlling the spread of this disease and minimizing its impact on the mango tree population cannot be overstated. The dataset contains 4000 photos separated into eight distinct classifications. Each class reflects a different type of mango leaf disease. The dataset has two types of data: training data and validation data. There are 3200 photos in the training data and 800 images in the validation data. This article investigated the significance of early detection of mango leaf disease and the possible application of the VGG 16 Transfer Learning Model to solve this issue. Our study in illness detection has achieved encouraging results through systematically collecting and preprocessing data, training a model with carefully selected hyperparameters, and evaluating the model's performance. Our work is very accurate, with a 97% accuracy rate. Using the VGG 16 Transfer Learning Model for mango leaf disease identification can drastically reduce the time and effort necessary for human inspection methods. This is especially crucial for large-scale agricultural companies. Aside from its multiple advantages, this technology has the potential for better precision and accuracy in assessing crop problems, allowing for enhanced farming practices and increased crop output.