Wheat | Precision Agriculture Lab

PrecisionAg Lab @ Uni.lu Workshop: Hyperspectral Remote Sensing & AI in Agriculture

Fri, 06 Feb 2026 00:00:00 +0000

The TUM Precision Agriculture Lab (PAG Lab) joined the scientific workshop “Hyperspectral Remote Sensing and AI in Agriculture” hosted by the University of Luxembourg (Campus Kirchberg, 4–5 February 2026). Our group delivered three talks connecting hyperspectral sensing, biophysical trait and SIF retrieval, and robust sensor calibration and AI pipelines for precision crop management.

Three Talks in the Precision Crop Management Session

Remote sensing of plant traits for nitrogen nutrition and yield estimation

Kang Yu gave a keynote talk and shared our lab’s work for trait-based crop monitoring that goes beyond generic vegetation indices—linking spectroscopy to interpretable variables relevant for nitrogen (N) nutrition, productivity, and management. We discussed how combining physics-informed approaches (e.g., radiative transfer concepts) with machine learning can improve generalization across environments and varieties, and help translate monitoring into practical recommendations and sensor development.

Assessing UAV-derived narrow-band SIF to characterize nitrogen use efficiency in wheat

Anirudh Belwalkar’s talk focused on narrow-band solar-induced fluorescence (SIF) from UAV platforms and its potential to characterize nitrogen use efficiency (NUE) signals in wheat. The talk connected physiological interpretation (photosynthetic functioning) with operational sensing constraints, discussing when UAV SIF can add value beyond reflectance-only products for N-related monitoring.

From reflectance to crop phenotype: quantifying uncertainty from UAV atmospheric correction methods

Wuhua Wang discussed the importance of hyperspectral data calibration and uncertainty estimates. This talk examined how different UAV atmospheric correction methods propagate uncertainty into downstream phenotyping and trait retrieval—a critical issue when hyperspectral workflows move from controlled experiments toward multi-site deployment.

Why This Workshop Mattered for Our Lab

The Uni.lu workshop brought together experts across hyperspectral sensing, UAV platforms, AI/ML, crop health, and nutrient strategies—creating a strong space for benchmarking methods, discussing data sharing and cross validation, and aligning on what “decision-ready” remote sensing should look like. For our PAG Lab, it was especially valuable to exchange views on Bridging physics + AI for stronger transferability (across years, sites, genotypes).

Thanks & Next Steps

We thank the University of Luxembourg, Prof. Teferle’s team and the workshop organisers for hosting a focused, high-quality event.

If you are interested in collaboration on hyperspectral trait retrieval, UAV SIF, or nitrogen decision support, feel free to reach out—we are keen to build shared datasets, benchmarks, and reproducible workflows. *— TUM Precision Agriculture Lab *

NH3-Min: Reducing NH₃ Losses from Application of Synthetic Nitrogen Fertilizers and Increasing Nitrogen Use Efficiency of Fertilization

Wed, 01 Apr 2020 00:00:00 +0000

Overview

NH3-Min is a national consortium project funded through the Landwirtschaftlichen Rentenbank, coordinated by the Thünen Institute for Climate-Smart Agriculture. The project quantifies and evaluates ammonia (NH₃) losses from synthetic nitrogen fertilizers and develops evidence-based mitigation measures, while simultaneously assessing nitrogen use efficiency (NUE) in arable crop production.

Around 15% of agricultural NH₃ emissions in Germany originate from synthetic N fertilizers. The four fertilizer types investigated—urea, calcium ammonium nitrate (CAN), ammonium nitrate urea solution (UAN), and ammonium sulfate urea—together account for more than 85% of NH₃ emissions from synthetic N fertilizers applied in Germany. Winter wheat serves as the model crop across all experimental sites.

Objectives

Quantify NH₃ emissions from different synthetic N fertilizer types and application systems across diverse German agro-climatic conditions.
Evaluate mitigation options including incorporation/injection techniques and urease inhibitors.
Derive site-differentiated NH₃ emission factors for synthetic N fertilizers under German production conditions.
Assess NUE of different synthetic N fertilizers and application strategies in field crop experiments.
Transfer knowledge to farmers, advisors, and policymakers.

Research Questions

How large are NH₃ emissions following application of different synthetic N fertilizers?
What are the site-differentiated NH₃ emission factors across Germany?
How effectively can application techniques (e.g., incorporation, injection) or inhibitors (urease inhibitors) reduce NH₃ emissions?
How do fertilizer-dependent NH₃ emissions affect yields, N uptake, and NUE in winter wheat?

Project Partners

Thünen Institute for Climate-Smart Agriculture
Technical University of Munich (TUM)
Julius Kühn-Institut – Federal Research Centre for Cultivated Plants (JKI)
KTBL – Association for Technology and Structures in Agriculture
Christian-Albrechts-Universität Kiel
Forschungszentrum Jülich
Technische Universität Berlin
Universität Hohenheim
Bayerische Landesanstalt für Landwirtschaft (LfL)
Ingenieurgemeinschaft für Landwirtschaft und Umwelt (IGLU)
Landwirtschaftskammer Niedersachsen
Stickstoffwerke Priesteritz (SKWP)

Duration: 01.2021 – 12.2024

GreenWindows4_0: Reduction of Greenhouse Gas Emissions and Ammonia by Optimized Nitrogen Management

Tue, 01 Jan 2019 00:00:00 +0000

Overview

GreenWindows4_0 is a research project funded by the Federal Ministry of Food and Agriculture (BMEL), jointly led by Prof. Urs Schmidhalter and Prof. Kang Yu at the Technical University of Munich (TUM). The project develops precision tools and workflows for optimizing nitrogen (N) fertilization in arable farming, with the dual goal of reducing greenhouse gas (GHG) emissions (e.g., N₂O) and ammonia (NH₃) volatilization while maintaining crop productivity and N use efficiency.

Objectives

Optimize N application timing and rates using satellite and UAV remote sensing to identify application windows that minimize emissions.
Monitor crop N status across field-scale experiments using hyperspectral and multispectral sensors.
Develop emissions-aware decision support linking remote sensing outputs to N management recommendations.
Improve N use efficiency (NUE) at the farm scale through data-driven approaches combining biophysical modelling and machine learning.

Methodology

The project integrates:

Satellite remote sensing (e.g., Sentinel-2) for large-area crop N status monitoring and yield estimation
UAV-based multispectral and hyperspectral imaging for high-resolution field phenotyping
Radiative transfer modelling and machine learning for robust trait retrieval
Field experiments with variable N fertilization treatments across multiple sites and seasons in Bavaria

Project Partners

Technical University of Munich (TUM)
AGRAVIS NetFarming GmbH
Funded by: Federal Ministry of Food and Agriculture (BMEL)

Duration: 01.01.2019 – 31.12.2022

Mokhtari et al. (2025). Satellite-based winter wheat yield estimation with a newly parameterized LUE model based on crop water status and leaf chlorophyll content. Field Crops Research. View

Wheat | Precision Agriculture Lab

PrecisionAg Lab @ Uni.lu Workshop: Hyperspectral Remote Sensing & AI in Agriculture

Three Talks in the Precision Crop Management Session

Remote sensing of plant traits for nitrogen nutrition and yield estimation

Assessing UAV-derived narrow-band SIF to characterize nitrogen use efficiency in wheat

From reflectance to crop phenotype: quantifying uncertainty from UAV atmospheric correction methods

Why This Workshop Mattered for Our Lab

Thanks & Next Steps

NH3-Min: Reducing NH₃ Losses from Application of Synthetic Nitrogen Fertilizers and Increasing Nitrogen Use Efficiency of Fertilization

Overview

Objectives

Research Questions

Project Partners

GreenWindows4_0: Reduction of Greenhouse Gas Emissions and Ammonia by Optimized Nitrogen Management

Overview

Objectives

Methodology

Project Partners

Related Publications