Diagnosis of Tomato Leaf Diseases Using Fractal Dimension and Box Counting Method Based on Image Analysis

Abstract

Diseases of tomato leaves, such as late blight, bacterial spot, and mosaic virus, pose a significant threat to agricultural production, necessitating early and accurate diagnosis. This study proposes a novel approach to diagnosing tomato leaf diseases based on fractal dimension analysis using the box-counting method. The method involves processing digital leaf images, binarization to highlight textural features, and calculating the fractal dimension, which reflects the complexity of the leaf structure. Experiments were conducted on a dataset of images comprising healthy leaves and leaves affected by various diseases to identify differences in fractal characteristics. The results demonstrate that the fractal dimension of healthy leaves significantly differs from that of diseased leaves and enables differentiation of disease types at early stages. The proposed approach shows potential for automating diagnostics in precision agriculture systems, offering high sensitivity and low implementation costs. Future research directions include integrating the method with machine learning technologies and developing mobile applications for real-time plant health monitoring.