AARTIFICIAL INTELLIGENCE IN ORGANIC AGRICULTURE — FROM PIXEL TO PLANT: AUTOMATED DETECTION OF PLANT DISEASES AND PESTS THROUGH MACHINE LEARNING TECHNIQUES

Abstract

Organic agriculture is a major challenge of the 21st century, requiring a balance between productivity and environmental sustainability. Within the EU’s Farm to Fork strategy, which aims for 25% organic farmland by 2030, early detection of plant diseases and pests is essential. Because synthetic pesticides are restricted, innovative solutions are needed, and artificial intelligence offers a promising alternative. This article examines computer vision and convolutional neural networks (CNNs) as tools for plant disease diagnosis. It analyzes architectures such as MobileNet, ResNet-50, YOLO, and EfficientNet in terms of performance and real-world applicability. It also highlights datasets like PlantVillage, widely used for AI training. The paper emphasizes accessible low-code platforms, including Teachable Machine and Roboflow, which enable educators and students to build models without advanced programming skills. Finally, it proposes the PlantGuard project, an interdisciplinary activity where students develop AI applications, fostering collaboration between informatics and biology while preparing learners for a digital and sustainable future.