A Comprehensive Survey on Deep Learning-based Techniques for Tomato Leaves Disease Detection and Classification

Abstract

Tomato is one of the most cultivated and consumed vegetable crops globally, but its yield is significantly threatened by a variety of diseases, primarily manifesting on the leaves. Early and accurate detection of these diseases is crucial for effective pest management and preventing substantial economic losses. Traditional methods, which rely on manual inspection by experts, are often slow, labor-intensive, and prone to human error. This survey paper provides a systematic and comprehensive review of the rapidly evolving field of automated tomato leaf disease detection, with a primary focus on deep learning (DL) techniques. We catalog a wide range of methodologies, from classical image processing and machine learning to state-of-the-art convolutional neural networks (CNNs) and vision transformers. The paper details publicly available datasets, discusses key technical challenges such as limited data, complex backgrounds, and real-time deployment, and analyzes the performance metrics of various approaches. Finally, we outline promising future research directions, including the integration of multimodal data, explainable AI (XAI), and the development of lightweight models for mobile and edge computing. This survey serves as a valuable resource for researchers and agricultural technologists aiming to understand the current landscape and contribute to advancing this critical application domain.

Keywords: deep learning, Convolutional Neural Networks (CNN), Image Classification, object detection, Tomato Leaf Disease, Plant Pathology, Vision Transformers, Precision Agriculture.