Conference Agenda

High-fidelity characterization of spray dynamics is a cornerstone of precision agriculture, yet it remains challenged by the intrusive nature and limited spatial resolution of traditional mechanical sampling. These conventional methods often fail to capture the complex spatial heterogeneity of atomization, masking critical details necessary for effective drift mitigation. To address this, this study proposes an atomization characteristics spatial distribution measurement framework based on phase Doppler interferometry (PDI) measurement technique to reconstruct the detailed spatial topology of spray fields. Unlike bulk statistical approaches, this fine-grained analysis characterized droplet velocity and size distributions with high fitting accuracy (R² > 0.90), revealing significant spatial heterogeneity. In nozzle-adjuvant system evaluations, the framework quantified how polymer adjuvants suppress liquid sheet instability, identifying optimal concentrations that reduce the volume fraction of drift-prone fine droplets (<150 μm) by over 45%. Crucially, a robust correlation was established between PDI-derived microscopic kinetic parameters and macroscopic drift percentages obtained from standardized wind tunnel tests. Consequently, a rapid drift prediction model was developed, offering an efficient potential alternative to time-consuming wind tunnel assessments. It establishes a novel research paradigm for the high-throughput screening of nozzle-adjuvant combinations and the rapid assessment of spray drift in agricultural engineering.

Tomato Leafminer (Phthorimaea absoluta) is a devastating invasive pest in China, threatening 11 vegetable families, particularly tomatoes. Current pheromone-based trapping equipment lacks intelligence and relies on manual inspection, resulting in low accuracy and slow response.

This study designed a low-power drive circuit based on the STM32 chip and developed intelligent camera control and image compression algorithms to capture real-time images of trapped Tomato Leafminer By integrating a 5G communication module, monitoring data is shared in real-time with a cloud platform. The device weighs ≤200 g and offers a battery life of ≥180 days on a single charge, enabling passive monitoring throughout the entire tomato planting cycle. Over 400 traps have been deployed in typical greenhouses, constructing a dataset comprising 20,000 trapping images. To address challenges such as low resolution of on-site trapping images, complex lighting conditions, and severe interference from color-trophic pests, we redesigned a pest detection model based on RT-DETR. By adding a small-object detection head, optimizing the backbone network, and introducing an efficient attention mechanism, the model's recognition accuracy for small-sized moth pests like Tomato Leafminer was significantly i

Peach palm (Bactris gasipaes), widely cultivated by smallholder farmers, presents management challenges due to the formation of dense and heterogeneous clumps, which hinder individual plant assessment and harvest planning. Unmanned aerial vehicle (UAV) multispectral imagery combined with vegetation index and structural canopy models offers a promising solution for canopy monitoring. This study developed a methodology to individualize peach palm clumps by integrating the Normalized Difference Red Edge Index (NDRE), the Canopy Height Model (CHM), and the watershed segmentation algorithm to quantify adult plants and support harvest planning. Multispectral images were acquired over a 0.4 ha peach palm area using a DJI Mavic 3M UAV and processed in Agisoft Metashape to generate the orthomosaic, Digital Surface Model (DSM), and Digital Terrain Model (DTM). The CHM was obtained by subtracting the DTM from the DSM, while NDRE was used to identify active vegetation. Crown segmentation was performed using the watershed algorithm, enabling clump-level individualization. Plants were classified as adult or young based on a 3 m height threshold. A total of 1,702 clumps were identified, including 281 adults (indicative of harvest) and 1,421 young plants. The methodology proved effective for plant-level segmentation and identifying adult plants with greater harvest potential.

Moringa oleifera Lam. is an emerging multipurpose crop increasingly introduced into Mediterranean agroecosystems due to its high nutritional value, rapid biomass production, and adaptability to semi-arid environments. Within precision agriculture frameworks, accurate monitoring of crop growth and spatial variability is essential to optimize irrigation management, biomass productivity, and seed yield. This study presents a two-year spatio-temporal evaluation of Moringa oleifera cultivation in Sicily through the integration of unmanned aerial vehicle (UAV) multispectral remote sensing and proximal field measurements. UAV surveys were conducted during key phenological stages to derive vegetation indices and canopy structural information, while simultaneous field measurements quantified biometric parameters including plant height, stem diameter, and biomass indicators. The integration of spectral and ground-based datasets enabled the characterization of crop vigor dynamics and intra-field variability across growing seasons. Statistical analyses revealed strong correlations between UAV-derived vegetation indices and biometric measurements (R² > 0.80), confirming the reliability of the proposed monitoring approach. Spatial variability patterns were associated with soil heterogeneity and irrigation management practices. These results demonstrate that the combined use of UAV and proximal sensing provides an effective and non-destructive method for monitoring Moringa oleifera growth and supports data-driven decision-making for sustainable crop management in Mediterranean precision agriculture systems.