Dataset validation for Disease Detection in Tomato Plants

Abstract

Tomato cultivation is a vital agricultural activity worldwide, contributing significantly to global food production. However, tomato crops are highly susceptible to various diseases, including mold, bacterial spot, and early blight, which can severely impact fruit quality and yield. These diseases, if not detected and managed promptly, lead to increased production costs and decreased efficiency. This research aims to address these challenges by developing and implementing an early disease detection dataset using Convolutional Neural Networks (CNNs). The system was trained with 4,083 images of tomato plants, allowing the CNN model to accurately identify specific diseases in both early and advanced stages. The model achieved a mean Average Precision (mAP) of 86.1%, a precision of 88.2%, and a recall of 82.6%, indicating its effectiveness of the dataset. This dataset can be used to develop different applications for managing tomatoes farm.