Publication

Abstract : Agriculture is crucial for society, by providing food, raw materials, and livelihoods. Timely detection and management of plant diseases are vital to minimize crop losses and maximize productivity. However, manual disease detection is time-consuming and subjective. This calls for automated solutions to plant health detection. Deep learning models have proved to give a promising solution to real-time image-based disease detection. There exist numerous research works that utilize these deep learning techniques to classify plant diseases. An efficient deep learning-based model is very important for the leaf disease classification problem. The work focusses on detecting and identification of disease in the leaf of tomato plant, one of the vegetables cultivated in the Ettimadai Village, Coimbatore, Tamil Nadu, India. The work employs convolutional neural network and VGG-16 that are trained, tested, and validated using available image samples. The models are fed with data available in Kaggle and tomato leaf images captured from J. P. Nursery, Mettankadu, Madukkarai, Tamil Nadu, India. The models are analysed for its performance and the best model is selected for real time prediction of diseases. The proposed work enhances disease identification, enables early detection, and potentially minimizes crop losses, thereby improving yields and income. The leveraging performance of artificial intelligence-based systems for plant disease detection contribute to food security, economic development, and sustainable farming practices.