Deep Learning for Coffee Leaf Diseases Detection in Precision Agriculture

Abstract

Coffee production faces challenges like climate change, drought, and biodiversity loss. Sustainable systems can improve crop yields and quality, but also threaten ecosystem function. AI can help classify and identify coffee leaf diseases, but traditional machine learning approaches struggle with big data. This study examines six deep learning models such as such as CNNs, ResNet50, MobileNet, GoogleNet, VGG16, and VGG19. The evaluation is done on the Kaggle dataset to classify between rust and miner diseases. MobileNet achieves superior results in terms of loss, accuracy, precision, recall, and F1-score with 0.0692, 0.973, 0.5625, 0.57143, 0.56693 respectively.