Harnessing Deep Learning for Disease Detection and Classification in Solanaceae Crops

As a vulnerable group to a host of diseases (fungal, bacterial or viral), Solanaceae crops such as tomatoes, potatoes and eggplants are economically vital food sources around the world. Diagnosis using traditional techniques is often labor intensive and not easily accessible to farmers in rural regions. Recent innovations in deep learning, along with the introduction of models like Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and their hybrid derivatives like CNN-LSTM, have facilitated robust automated imaging-based plant disease detection systems. Such models provide high classification accuracy and automate disease diagnostic processes, minimizing dependence on manual work. However, challenges like scarcity of data, noise within environmental surroundings, visual likeness of symptomatic lesions, and interpretable models still prevail. In this chapter, we analyse advanced deep learning techniques applied to disease detection in Solanaceae plants. We discuss the overall model architecture, the data augmentation strategies used, and the evaluation metrics adopted to assess performance. The chapter also highlights current research limits and emerging directions such as multi-modal and explainable AI, few-shot learning, edge computing, model-agnostic approaches, and improved salient feature based frameworks. These AI-based precision agricultural control systems, coupled with ready-for-field deployment technologies, could dramatically bolster sustainability, food security, and resilience of crops.