A Novel and Efficient Deep Learning Models for Assessing AI’s Impact on Disease Diagnosis in Agriculture

Background: Agriculture sustains human life by providing food, raw materials and employment opportunities. However, climate change and resource limitations pose significant challenges to crop production. AI-driven smart farming has emerged as a solution to enhance agricultural efficiency, with Explainable AI (XAI) improving transparency in decision-making. Innovations such as smart sensors and automated systems have benefited key agricultural sectors, including crops, forestry, livestock and aquaculture. Turmeric, valued for its medicinal and economic significance, requires careful monitoring to combat diseases like Leaf Spot and Leaf Blotch, which can impact yield and quality. Methods: This study introduces Turmeric Net, a Convolutional Neural Network (CNN)-based model leveraging transfer learning to detect and classify turmeric leaf diseases. The dataset used consists of 791 original images and 3,702 augmented images obtained from Mendeley Data, categorized into four classes: healthy leaf, dry leaf, leaf blotch and rhizome rot. The model development was carried out using TensorFlow, with ResNet50V2 as a baseline for comparison. The models were trained on processed image data, incorporating augmentation techniques to improve robustness and generalizability. Result: The accuracy of both models was evaluated. ResNet50V2 achieved an accuracy exceeding 99%, demonstrating high effectiveness in disease classification. Meanwhile, TurmericNet attained a competitive accuracy of 98%, making it a reliable alternative for turmeric disease identification. These results indicate that deep learning-based models can significantly aid in early disease detection, providing farmers with a valuable tool to enhance crop management and productivity.