GIS Day at FAZOS: workshop “The life story of maize, seen from space”

The AgroVeFe project marks this year’s GIS Day (GIS Day 2026) at the Faculty of Agrobiotechnical Sciences Osijek with a talk and workshop, “The life story of maize, seen from space”. The workshop takes place on Friday, 20 November 2026, starting at 9:00 in the Faculty Council Hall (1st floor), and is led by Lucija Galić, PhD, from the Department of Geoinformation Technologies and GIS, a member of the project research team.

The topic is how the Sentinel-2 satellite of the European Copernicus programme follows a crop from sowing to harvest. Using production plots at the FAZOS experimental station in Tenja, participants will see how fields “light up” and “fade out” over the season, what the NDVI curve tells us about the crop, and why such data matter for predicting maize yield – the main goal of the AgroVeFe project.

NDVI, the Normalized Difference Vegetation Index, is calculated from the satellite’s near-infrared (B8) and red (B4) bands: NDVI = (B8 − B4) / (B8 + B4). The more vigorous and healthy the crop, the higher the value, and its change over the season reveals when the maize emerged, when it reached peak development and when it began to mature.

At a glance

  • What: talk and workshop “The life story of maize, seen from space” (in Croatian)
  • When: Friday, 20 November 2026, 9:00
  • Where: Faculty Council Hall (1st floor), Faculty of Agrobiotechnical Sciences Osijek, Vladimira Preloga 1, Osijek
  • Led by: Lucija Galić, PhD, Department of Geoinformation Technologies and GIS
  • For: FAZOS staff and students

The workshop is one of the science outreach activities of the AgroVeFe project. Read more about how a satellite “sees” maize in How does a satellite see maize?

Satellite images on the poster: Copernicus Sentinel-2 (2024), ESA; processing: Copernicus Browser.

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

First year of AgroVeFe: 11 scientific papers, including a Q1 paper on the project methodology

Ilustracija: metlice kukuruza u sumrak pod večernjim nebom

In the first year of the AgroVeFe project (October 2025 – September 2026), members of the research team published 11 scientific papers acknowledging the project’s support: 10 in journals indexed in Web of Science and Scopus (6 WoSCC-SCI-Expanded, 4 WoSCC-ESCI) and one original scientific paper in the proceedings of an international conference. Nine papers are available in open access.

We particularly highlight the paper by Radočaj, Jurišić, Plaščak and Galić in the Q1 journal International Journal of Remote Sensing. It shows how within-field maize yield can be predicted from Sentinel-2 phenological metrics and an interpretable stacked ensemble machine learning model, and it is the methodological basis for yield potential modelling in the second year of the project: doi.org/10.1080/01431161.2026.2709705.

Also directly related to the project objectives are papers comparing ordinary and spatial cross-validation in maize yield prediction (Eng 6(10) 270), a review of satellite methods for yield prediction (Agriculture 16(4) 417) and a bibliometric analysis of machine and deep learning in remote sensing for precision agriculture (Agronomy 16(8) 807).

The full list of papers is on the Publications page.

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

AgroVeFe establishes scientific cooperation with Belgrade and Banja Luka

The AgroVeFe project has signed two scientific research cooperation agreements: with the Faculty of Agriculture, University of Belgrade (Serbia), and with the Agricultural Institute of Republic of Srpska in Banja Luka (Bosnia and Herzegovina).

With the Belgrade partner, mentored by Prof. Miloš Pajić, PhD, work will focus on integrating and validating ISOXML maps in the control systems of agricultural machinery, where Assist. Prof. Željko Barač, PhD, will undertake a research stay. In Banja Luka, mentored by Tihomir Predić, PhD, Assist. Prof. Dorijan Radočaj, PhD, will train in calibrating machine learning models on Sentinel-2 data.

The cooperation includes exchanging data and methodologies, joint scientific papers and project proposals, and validating the models under the different agro-ecological conditions of Serbia and Bosnia and Herzegovina.

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

Soil electrical conductivity scanning system with RTK GNSS acquired

As part of the AgroVeFe project, the Faculty of Agrobiotechnical Sciences Osijek has acquired a Geonics EM38-4 system for non-contact scanning of soil electrical conductivity (EC), with an RTK GNSS receiver, field controller and software. The device has been delivered and put into operation, and team members completed eight hours of field training.

Soil electrical conductivity reveals differences in texture, moisture and salt content within a field, while RTK GNSS records the position of each measurement with a precision of about 2 cm. In the project, EC maps will be compared with yield potential predictions from satellite imagery to determine whether combining soil and satellite data gives more reliable zones for variable-rate application of inputs.

Audio-video equipment with editing software was also purchased to produce educational videos for students and farmers. All equipment purchased with project funds is labelled “Funded by the European Union – NextGenerationEU”.

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

How does a satellite see maize?

Ilustracija: pogled iz zraka na mozaik poljoprivrednih parcela

The Sentinel-2 satellite of the European Copernicus programme images every point in Croatia every few days, and the images are free for everyone. From the ratio of reflected red and near-infrared light, vegetation indices such as NDVI are calculated, showing how green and vigorous a crop is.

The problem is that at the peak of maize growth, classic indices “saturate” and stop distinguishing a good crop from an excellent one. That is why, in addition to NDVI, EVI, EVI2 and SAVI, the AgroVeFe project also uses four saturation-resistant indices: WDRVI, PPI, IDVI and NDVI3RE.

We also track crop phenology: when the crop emerges, when it reaches peak development and when it begins to senesce. Seven phenological metrics give context to the spectral data, so the model knows which growth stage the plant is in when the satellite takes an image.

In the first year of the project, the team set up automated downloading of Sentinel-2 imagery for 2020–2025 in Google Earth Engine, removal of clouds and shadows, and harmonisation of images to a spatial resolution of 20 m for three pilot sites in eastern Croatia. The availability of cloud-free images during the maize growing season proved satisfactory.

In the second year, the team will calculate all 56 combinations of indices and phenological metrics for each reference point and use them to train the first machine learning models.

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

Call to farms: become a partner of the AgroVeFe project

Ilustracija: pogled iz zraka na poljoprivredne parcele u kasno ljeto

The AgroVeFe project is looking for farms that would like to take part in research on maize yield potential prediction.

Who we are looking for

  • farms with combine harvesters equipped with a yield sensor that can share maize yield data from previous seasons
  • farms that can provide a field for testing soil electrical conductivity scanning and zone maps

What partners get

  • yield potential maps of their fields derived from satellite imagery
  • soil electrical conductivity maps with RTK GNSS precision
  • first ISOXML zone maps for variable-rate seeding and fertilisation, ready to load into agricultural machinery
  • invitations to project workshops and demonstrations

Shared data is used solely for research purposes, under a signed cooperation and confidentiality agreement.

Contact: Prof. Ivan Plaščak, PhD, iplascak@fazos.hr

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

AgroVeFe project kick-off conference held

Početna konferencija projekta AgroVeFe, FAZOS, 26. veljače 2026.

The kick-off conference of the AgroVeFe project was held on 26 February 2026 at the Faculty of Agrobiotechnical Sciences Osijek. The research team presented the project objectives, methodology, eight work packages and timeline to researchers, students and representatives of farms.

Project leader Prof. Ivan Plaščak highlighted the five research gaps the project addresses: linking saturation-resistant vegetation indices with phenological metrics, multi-year model validation, the inclusion of soil properties, making precision agriculture accessible to small farms, and outputs that can be applied in the field. The team will test 56 combinations of eight vegetation indices and seven phenological metrics on yield data from 2020–2025.

The conference opened discussions with farm partners about sharing data and future testing of the solution on their fields.

Download the conference presentation (PDF, in Croatian) · All photos in the gallery

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

First AgroVeFe research team meeting held

Ilustracija: satelit snima polje kukuruza

On 20 November 2025, the first official meeting of the AgroVeFe research team was held in Osijek. The meeting launched project implementation after the contract was signed.

Project leader Prof. Ivan Plaščak presented the context and objectives of the research, an overview of previous research on yield prediction from satellite data and the five research gaps the project addresses. The team went through the project methodology: from collecting maize yield data for 2020–2025 and Sentinel-2 time series, through calculating 56 combinations of vegetation indices and phenological metrics, to machine learning models, ISOXML zone maps and soil electrical conductivity scanning.

The team agreed on roles in the eight work packages, a year-by-year timeline and a risk matrix with mitigation measures. It committed to monthly meetings and quarterly progress reviews per work package, while scripts, data and reports are kept in a shared repository with version control on GitHub. The visibility rules for European Union – NextGenerationEU funding, which apply to all papers, posters and project communications, were also presented.

Download the presentation from the first research team meeting (PDF, in Croatian)

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.

AgroVeFe project launched: satellite imagery and machine learning for maize yield prediction

On 3 November 2025, the Faculty of Agrobiotechnical Sciences Osijek (FAZOS) signed the contract for the institutional research project AgroVeFe – Prediction of maize yield potential using machine learning models based on vegetation indices and phenological metrics from Sentinel-2 multispectral satellite imagery (contract no. 581-UNIOS-30).

The project runs for 48 months, has a budget of EUR 130,000 and is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan. It is led by Prof. Ivan Plaščak, PhD, Head of the Department of Agricultural Engineering and Renewable Energy Sources.

Precision agriculture in Croatia is often held back by the cost of sensors and yield mapping systems. AgroVeFe investigates whether free Sentinel-2 satellite imagery, combined with machine learning models and soil electrical conductivity measurements, can provide a reliable yield potential map. The result will be zone maps in ISOXML format that agricultural machinery can load directly for variable-rate seeding and fertilisation.

What the project aims to achieve

Scientific results

  • at least 15 scientific papers in journals indexed in Web of Science and Scopus (Q1/Q2)
  • at least 50 spectral-phenological covariates for yield potential prediction (8 vegetation indices × 7 phenological metrics)
  • yield potential prediction accuracy expressed as a coefficient of determination R² ≥ 0.7
  • a machine learning model validated on at least 100 ha of agricultural land
  • interdisciplinary research across three scientific fields

Open science

  • 3 publicly available datasets
  • 2 open-source digital tools for data analysis (GEE and R scripts)
  • 2 measures to promote open science

Cooperation and practical application

  • 5 formal partnerships with farms
  • 1 patent application
  • 2 collaborations with international scientific partners
  • 2 researchers in international mobility and 2 teachers in international exchange

Outreach and infrastructure

  • 11 science outreach activities, including workshops for students and farmers, three events and educational videos
  • equipping the laboratory and field team: a soil electrical conductivity scanning system with RTK GNSS, a drone with a multispectral camera and GIS workstations
  • an investment of EUR 20,000 of FAZOS’s own funds in research equipment

AgroVeFe (581-UNIOS-30) is funded by the European Union – NextGenerationEU through the National Recovery and Resilience Plan 2021–2026.