Abvstract:
Water scarcity and climate change necessitate a paradigm shift towards increasingly predictive and sustainable precision agriculture. In this context, this seminar illustrates research regarding the application of Artificial Intelligence (AI) for real-time monitoring of plant physiology, aiming to optimize water usage and detect early crop stress, with a specific focus on tomato plants.
At the core of this work is the analysis of data from the “Bioristor", an organic electrochemical transistor (OECT) inserted directly into the plant stem. Unlike traditional instrumentation, this device continuously monitors ion concentration within the sap, offering a unique "internal" perspective on plant health.
The presentation will illustrate how advanced Machine Learning techniques have been employed to decode these bio-electronic signals. We will discuss the challenge of modeling evolving biological systems and how the adoption of adaptive approaches (such as Fuzzy Decision Trees and Incremental Learning) has enabled the development of accurate and interpretable predictive models.
Finally, experimental results confirming the models' ability to dynamically adapt to the plant's physiological evolution will be presented, concluding with recent developments towards a multimodal approach that fuses Bioristor signals with computer vision for more comprehensive phenotyping
Login [pagine riservate]
Intranet
Webmail
LinkedIn