Privacy-Preserving Soil Data: Federated Learning for Topsoil Descriptors Via Remote Sensing

Nikolaos Tsakiridis; Eleni Kalopesa; Nikolaos Tziolas; Nick Berkvens; Theodoros Chalazas; Tuna Coppens; Panagiotis Ilias

doi:10.1109/IGARSS55030.2025.11242409

Privacy-Preserving Soil Data: Federated Learning for Topsoil Descriptors Via Remote Sensing

Nikolaos Tsakiridis
, Eleni Kalopesa
, Nikolaos Tziolas
, Nick Berkvens
, Theodoros Chalazas
, Tuna Coppens
, Panagiotis Ilias

Technology and Food Science

Research output: Contribution to conference › C3: Conference - meeting abstract › peer-review

Abstract

This paper explores the application of federated learning in soil science to address the challenges of data privacy and collaboration. Leveraging decentralized machine learning, we demonstrate how to predict soil properties effectively while preserving privacy. Our study focuses on cropland data from Flanders, Belgium, and Central Macedonia, Greece, and utilizes the LUCAS topsoil database and Copernicus Sentinel-2 data as an initial model. The results showcase the potential of federated learning combined with Earth observation to advance soil monitoring and sustainable land management.

Original language	English
DOIs	https://doi.org/10.1109/IGARSS55030.2025.11242409
Publication status	Published - 5-Aug-2025
Event	IEEE International Geoscience and Remote Sensing Symposium 2025 - Brisbane, Australia Duration: 3-Aug-2025 → 8-Aug-2025 https://2025.ieeeigarss.org/

Conference

Conference	IEEE International Geoscience and Remote Sensing Symposium 2025
Country/Territory	Australia
City	Brisbane
Period	3/08/25 → 8/08/25
Internet address	https://2025.ieeeigarss.org/

Access to Document

10.1109/IGARSS55030.2025.11242409

SCALEAGDATA: Scaling agricultural sensor data for an improved monitoring of agri-environmental conditions
Vangeyte, J. (ProjectSupervisor), Berkvens, N. (Researcher), Coppens, T. (Project Manager), Chalazas, T. (Researcher), Van Loo, K. (Researcher), Saberioon, M. (Former Researcher), Ilias, P. (Former Researcher), De Man, W. (Researcher) & Bauwens, J. (Researcher)
1/01/23 → 31/12/26
Project: Research

Cite this

@conference{281f84c720ec40bdb756717b3674e457,

title = "Privacy-Preserving Soil Data: Federated Learning for Topsoil Descriptors Via Remote Sensing",

abstract = "This paper explores the application of federated learning in soil science to address the challenges of data privacy and collaboration. Leveraging decentralized machine learning, we demonstrate how to predict soil properties effectively while preserving privacy. Our study focuses on cropland data from Flanders, Belgium, and Central Macedonia, Greece, and utilizes the LUCAS topsoil database and Copernicus Sentinel-2 data as an initial model. The results showcase the potential of federated learning combined with Earth observation to advance soil monitoring and sustainable land management.",

author = "Nikolaos Tsakiridis and Eleni Kalopesa and Nikolaos Tziolas and Nick Berkvens and Theodoros Chalazas and Tuna Coppens and Panagiotis Ilias",

year = "2025",

month = aug,

day = "5",

doi = "10.1109/IGARSS55030.2025.11242409",

language = "English",

note = "IEEE International Geoscience and Remote Sensing Symposium 2025 ; Conference date: 03-08-2025 Through 08-08-2025",

url = "https://2025.ieeeigarss.org/",

}

TY - CONF

T1 - Privacy-Preserving Soil Data: Federated Learning for Topsoil Descriptors Via Remote Sensing

AU - Tsakiridis, Nikolaos

AU - Kalopesa, Eleni

AU - Tziolas, Nikolaos

AU - Berkvens, Nick

AU - Chalazas, Theodoros

AU - Coppens, Tuna

AU - Ilias, Panagiotis

PY - 2025/8/5

Y1 - 2025/8/5

N2 - This paper explores the application of federated learning in soil science to address the challenges of data privacy and collaboration. Leveraging decentralized machine learning, we demonstrate how to predict soil properties effectively while preserving privacy. Our study focuses on cropland data from Flanders, Belgium, and Central Macedonia, Greece, and utilizes the LUCAS topsoil database and Copernicus Sentinel-2 data as an initial model. The results showcase the potential of federated learning combined with Earth observation to advance soil monitoring and sustainable land management.

AB - This paper explores the application of federated learning in soil science to address the challenges of data privacy and collaboration. Leveraging decentralized machine learning, we demonstrate how to predict soil properties effectively while preserving privacy. Our study focuses on cropland data from Flanders, Belgium, and Central Macedonia, Greece, and utilizes the LUCAS topsoil database and Copernicus Sentinel-2 data as an initial model. The results showcase the potential of federated learning combined with Earth observation to advance soil monitoring and sustainable land management.

U2 - 10.1109/IGARSS55030.2025.11242409

DO - 10.1109/IGARSS55030.2025.11242409

M3 - C3: Conference - meeting abstract

T2 - IEEE International Geoscience and Remote Sensing Symposium 2025

Y2 - 3 August 2025 through 8 August 2025

ER -

Privacy-Preserving Soil Data: Federated Learning for Topsoil Descriptors Via Remote Sensing

Abstract

Conference

Access to Document

Fingerprint

Projects

SCALEAGDATA: Scaling agricultural sensor data for an improved monitoring of agri-environmental conditions

Cite this