GraDA: an integrated system to ensure quality in table grapes

Table grape production and quality is highly influenced by environmental conditions, with abrupt weather changes increasing the risk of vineyard diseases. The integration of smart agriculture technologies enables early detection and effective management of these threats, improving crop quality and yield. This work proposes the GraDA system, a data collection and grape disease risk assessment system, consisting of three main components, namely the GraDA platform, a mobile application and an AI-driven grape disease risk prediction model. The mobile application is responsible for collecting images of grapes and leaves, while the AI-driven disease risk prediction model processes multi-modal input data—grape images, leaf images, and meteorological data—to achieve enhanced grape disease risk prediction accuracy. Finally, the GraDA platform is concerned with the collection and visualization of imaging and meteorological data through a web-based interface, enabling agronomists and farmers to monitor in real-time the weather conditions and the estimated disease risk scores, as well as manage farming activities, thus facilitating data-driven decision-making for vineyard health optimization. The proposed AI model’s accuracy was evaluated in a multi-modal dataset, achieving a performance above 93% in grape disease risk estimation, while the usability of the overall GraDA system was assessed through questionnaires distributed to agronomists and farmers, leading to a score of 79.83% in the SUS scale, indicating excellent usability and user satisfaction from the proposed system. Finally, a pilot study of the GraDA system was performed to assess the applicability of the system in real-life scenarios.