PROJECT OVERVIEW

The MUSELY project aims to establish an innovative framework for diagnosing plant diseases and monitoring vegetation responses to pathogens through the integration of hyperspectral imaging, Solar-Induced Fluorescence (SIF), and thermal-infrared (TIR) and LiDAR data. The multi-sensor remote sensing approach provides new and detailed insights into plant structural biochemical and physiological responses across disease temporal phases: (i) early physiological indicators for pre-symptomatic detection, (ii) diseases-specific spectral indicators for identifying and quantifying visible symptoms, and (iii) long-term monitoring of structural and biochemical changes to assess resilience.

The project employs exemplary case studies based on controlled field experiments and multi-sensor Remote Sensing (RS) observations using Unmanned Aerial Vehicles (UAVs) to develop and test innovative methodologies and approaches. Specifically, Fusarium Head Blight in wheat (Experiment I), Diaporthe helianthi infection in sunflower (Experiment II), and Fusarium wilt in tomatoes (Experiment III) are chosen as key case studies, as they represent major diseases with distinct symptoms and pathogenic mechanisms. Complementary leaf level physiological measurements further support the assessment and interpretation of remote sensing data, while an analysis of root-beneficial microbiota provides additional pioneering insights.

Field experiments are conducted to advance RS methodologies across plots of varying cultivar susceptibility and pathogen treatments. Multiscale RS data are collected with UAV-based spectral imaging over the entire field, complemented by ground-based top-of-canopy and leaf-level field spectroscopy and SIF measurements. The MUSELY project aims to develop an advanced, high-throughput framework for diagnosing plant diseases and monitoring vegetation responses to pathogens at various stages of disease progression. The key objectives include refining data collection protocols across scales from leaf to UAV, utilizing SIF and TIR indicators for early disease detection, and exploring hyperspectral features to identify specific plant pathogens. The project will also assess methods for quantifying plant leaf and canopy indicators using advanced machine learning algorithms, while evaluating the economic and environmental benefits of the technical approach proposed.

The project leverages expertise from the University of Milan Bicocca (UNIMIB), the University of Pisa (UNIPI), and the National Research Council (CNR) across disciplines including remote sensing, plant pathology, microbiology and agricultural economics. The project is organized into five work packages (WPs) addressing a specific set of activities (Figure 1). WP1 focuses on project management, reporting, and dissemination to facilitate coordination and maximize impact. WP2 is dedicated to setting up and characterizing field experiments, including plant species selection and site preparation. WP3 covers remote sensing, ground-based data acquisition to capture multi-scale data on vegetation responses. WP4 centers on data analysis and model machine learning development for disease diagnostics, while WP 5 assesses the economic and environmental benefits of the technologies, particularly for early plant disease detection.