A Review on Automated Tomato Leaf Disease Detection Using Deep Learning in Smart Farming Systems

Abstract

Tomato cultivation is an important agricultural practice across the world, but it confronts many obstacles because of many diseases that negatively affect crop quality and production. This article discusses deep learning-based automated tomato leaf disease sensors for smart farming systems. It discusses publicly accessible datasets such as PlantVillage and other large repositories, as well as preprocessing and data augmentation techniques to improve model performance. Different DL networks, such as Convolutional Neural Network (CNNs), MobileNet, VGG, ResNet, Inception and YOLO, are analyzed based on their soundness in classification and detection. Besides, the review has indicated the combination of artificial intelligence and internet of things, drones and cloud-based decision support systems to provide real-time monitoring of crops. Critical issues like imbalance in the data, complexity of computation, inadequate field data and unexplainability are also discussed. Lastly, the research directions in the future, including lightweight models, edge computing and explainable AI, are discussed in order to move towards more accurate solutions of precision agriculture and sustainable smart farming.