Self-attention and frequency-augmentation for unsupervised domain adaptation in satellite image-based time series classification

Wed, 01 Jan 2025 00:00:00 +0000

David Gackstetter, Kang Yu, & Marco Körner (2025). Self-attention and frequency-augmentation for unsupervised domain adaptation in satellite image-based time series classification. ISPRS Journal of Photogrammetry and Remote Sensing, 224: 113–132.

Approaching holistic crop type mapping in Europe through winter vegetation classification and the Hierarchical Crop and Agriculture Taxonomy

Mon, 01 Jan 2024 00:00:00 +0000

David Gackstetter, Marco Körner, & Kang Yu (2024). Approaching holistic crop type mapping in Europe through winter vegetation classification and the Hierarchical Crop and Agriculture Taxonomy. International Journal of Applied Earth Observation and Geoinformation, 134: 104159.

DigiCrop Farm Data: Precision Farming Data Analytics at TUM Farm

Sun, 01 May 2022 00:00:00 +0000

Overview

DigiCrop Farm Data is a collaborative project funded through the Hans Eisenmann-Forum für Agrarwissenschaften (HEF) at TUM, jointly led by Prof. Kang Yu (Precision Agriculture), Prof. Heinz Bernhardt (Agricultural Systems Engineering), and Prof. Timo Oksanen (Agricultural Robotics and Automation). The project develops precision farming data analytics workflows using multi-sensor UAV and field datasets collected at the TUM Dürnast research farm, bridging agronomy, sensing, robotics, and data science.

Objectives

Integrate multi-sensor UAV data (multispectral, RGB, LiDAR) with ground-based farm records for comprehensive field-scale crop monitoring.
Develop data analytics pipelines for extracting agronomically meaningful insights from heterogeneous precision farming datasets.
Assess crop nitrogen status and NUE using drone remote sensing and field measurements at TUM’s experimental farm.
Support farm management decisions by linking remote sensing outputs to fertilization and yield outcomes.

Methodology

The project combines:

Cross-disciplinary collaboration between precision agriculture, agricultural engineering, agrimechatronics and robotics groups at TUM
Digitize Field experiment data to enalbe reuse of historical farming data e.g., for enhancing variable nitrogen fertilization applications
Data analytics and machine learning for trait retrieval, canopy characterization, and NUE assessment
Mobile sensing and remote sensing data calibration at TUM’s Dürnast research farm for spatial crop monitoring
Funded by: Hans Eisenmann-Forum für Agrarwissenschaften (HEF) Seed Fund 2022

Duration: 5/2022 – 5/2023

Wang et al. (2025). Drone multispectral imaging captures the effects of soil mineral nitrogen on canopy structure and nitrogen use efficiency in wheat. Computers and Electronics in Agriculture. View