Preunec GmbH

10/10/2026

Low NDVI doesn't always mean a sick crop. Sometimes the wheat is just ripe. 🌾

Software 2 is SMART Farming's Autonomous Decision-Making Model: it combines Software 1's nitrogen and irrigation scores (Low / Medium / Ideal) with the number of days after sowing (DAS) and gives the farmer a recommendation.

The plant's age is translated into a growth stage using the international Zadoks growth scale and NDVI (Normalized Difference Vegetation Index) data from the Adana field. Each plot's NDVI is then checked against what is normal for that stage, so the same score turns into different advice depending on the stage.

The recommendation is written by the Llama 3.3 70B language model, because it can weigh the scores in context and account for trade-offs such as the nitrogen–water balance.

Software 2 was developed by preunec.

Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Education and Culture Executive Agency (EACEA). Neither the European Union nor EACEA can be held responsible for them.

07/10/2026

03/10/2026

We tested 21 AI models for SMART Farming's first software — and didn't pick the most accurate one.
Software 1 uses drone NDVI (Normalized Difference Vegetation Index) data from Çukurova University's 54 wheat plots in Adana to estimate a plot's nitrogen and irrigation level as Low / Medium / Ideal, and those estimates are the input for Software 2. The model wasn't chosen on theory: 12 machine-learning models and 9 large language models were tested on the project's data. The LLMs came out 5–7 points ahead, but at an API cost for every prediction. Naive Bayes was chosen for nitrogen and Decision Tree for irrigation: they run locally and give the same result every time. The research and Software 1 were developed by preunec.
smartfarm-ai.com/en
preunec.com/en

Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Education and Culture Executive Agency (EACEA). Neither the European Union nor EACEA can be held responsible for them.

26/09/2026

In Arad, the project grows wheat in pots — on purpose.

"Aurel Vlaicu" University of Arad, which leads the project's research, analysis and evaluation work package, studies wheat in a controlled pot trial. The soil came from a farm and the pots were kept in the open air. There are three stress factors: soil humidity, nitrogen, and the cold that comes with sowing a month late. Each is compared against a control group.

The log from November to April: slow development, a sharp cold snap, and cold overlapping with a lack of water in the soil.

The April interventions: differentiated fertilisation, photosynthesis measurements, spring wheat sown, germination in four days.



Funded by the European Union.

19/09/2026

Before the first drone flew, a field in Adana spent five months getting ready for it.

Çukurova University's Field Crops and Landscape Architecture departments set up the field that the project's drone imagery comes from. The work began with a visit to the Eastern Mediterranean Agricultural Research Institute, an associate partner of the project.

The site had to pass two sets of criteria, one for the drones and one for the crop, with ownership/permissions and technical support for the hardware also in place, and about one hectare was selected. Silage corn was planted first to draw down the nitrogen left over from the previous season, so the nitrogen trial could start from a clean slate.

In December 2024 the block plan was applied: 54 plots across two blocks, nine nitrogen × irrigation combinations, early and late sowing, with TSP fertiliser at sowing.



Funded by the European Union.

16/09/2026

What does smart farming mean to an organic farmer?

Our partner VondiConsulting asked an organic farmer four questions about farming and technology. One answer per video, every Wednesday.

For this farmer, smart farming means protecting resources and making full use of the growing season: planting crops with low and high nutrient needs side by side, and mulching the soil to save water and increase its biodiversity.

Next Wednesday: what are the biggest challenges in farming?

smartfarm-ai.com/en



Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Education and Culture Executive Agency (EACEA). Neither the European Union nor EACEA can be held responsible for them.

12/09/2026

Lernen Sie SMART Farmings Partner in Adana kennen: die Universität Çukurova.
Mit Sitz in Adana trägt die Universität Çukurova mit ihrem Forschungshintergrund in Landwirtschaft und Geoinformationstechnologien zum Projekt SMART Farming bei. Ihre Fakultät für Landwirtschaft und Fakultät für Architektur leiten über das Institut für Feldpflanzenbau und das Institut für Landschaftsarchitektur die Datenerfassung und -verarbeitung im Rahmen von Evaluation II. Über zwei Jahre und auf 54 Versuchsparzellen untersuchen drohnengestützte Versuche Trockenstress und Stickstoffmangel mittels NDVI. Die Universität trägt zudem zu Seminaren und Multiplikatorenveranstaltungen in Adana bei.

Finanziert von der Europäischen Union. Die geäußerten Ansichten und Meinungen entsprechen jedoch ausschließlich denen der Verfasser und geben nicht unbedingt die Ansichten der Europäischen Union oder der Europäischen Exekutivagentur für Bildung und Kultur (EACEA) wieder. Weder die Europäische Union noch die EACEA können dafür verantwortlich gemacht werden.

12/09/2026

Faceți cunoștință cu partenerul SMART Farming din Adana: Universitatea Çukurova.
Cu sediul în Adana, Universitatea Çukurova contribuie la proiectul SMART Farming prin experiența sa de cercetare în agricultură și tehnologii geospațiale. Facultatea de Agricultură și Facultatea de Arhitectură, prin Departamentul de Culturi de Câmp și Departamentul de Arhitectură Peisagistică, conduc colectarea și procesarea datelor în cadrul Evaluării II. Timp de doi ani și pe 54 de parcele experimentale, testele cu drone examinează stresul hidric și deficitul de azot prin NDVI. Universitatea contribuie, de asemenea, la seminarii și evenimente multiplicatoare în Adana.

Finanțat de Uniunea Europeană. Punctele de vedere și opiniile exprimate aparțin, însă, exclusiv autorului (autorilor) și nu reflectă neapărat punctele de vedere și opiniile Uniunii Europene sau ale Agenției Executive Europene pentru Educație și Cultură (EACEA). Nici Uniunea Europeană și nici EACEA nu pot fi făcute responsabile pentru acestea.