SCI for Sustainable Sugar
Revolutionizing sugar-beet production through AI-driven precision agriculture
Learn More
A Collaborative Vision for Digital Agriculture
SCI for Sustainable Sugar represents a groundbreaking ICT-AGRI-FOOD 2022 research initiative that brings together leading experts in agricultural technology, industry operations, and scientific research. Coordinated by Agrovisio, this consortium unites the practical expertise of Kayseri Sugar Factory with the research excellence of Ege University and the irrigation innovation of Rivulis.
The project's ambitious goal is to fundamentally transform sugar-beet production from a traditional agricultural practice into a data-driven, sustainable enterprise. By integrating multiple data sources—satellite imagery, ground sensors, drone surveillance, and laboratory analysis—into a unified AI-powered platform, the team has created a comprehensive decision-support system that empowers farmers and factory managers to make informed, real-time decisions about crop health, soil management, and resource optimization.
Agrovisio
Project coordinator and platform developer
Kayseri Sugar Factory
Industry partner and pilot site host
Ege University
Research and validation expertise
Rivulis
Irrigation technology specialist
Building a Comprehensive Agricultural Data Ecosystem
Between 2023 and 2025, the SCI consortium constructed a sophisticated data-space architecture that serves as the backbone of the entire decision-support system. This architecture seamlessly integrates multiple data streams, creating a holistic view of crop conditions and environmental factors that influence sugar-beet production.
The system combines high-resolution Sentinel-2 and PlanetScope satellite imagery with ground-based IoT soil-moisture sensors, comprehensive meteorological data from local weather stations, and detailed laboratory analyses of field samples. This multi-layered approach ensures that farmers and agronomists have access to both the broad spatial perspective provided by satellite monitoring and the precise, localized insights from field sensors and laboratory testing.
01
Satellite Data Collection
Continuous monitoring via Sentinel-2 and PlanetScope for vegetation indices and crop health assessment
02
Ground Sensor Integration
Real-time soil moisture, temperature, and environmental data from IoT sensor networks
03
Laboratory Analysis
Detailed field-sample testing for nutrient content, soil composition, and quality parameters
04
Data Fusion & Processing
AI-powered integration of all data streams into actionable insights and predictions
Machine Learning Models for Precision Prediction
At the heart of the SCI platform lies a suite of advanced machine-learning models specifically designed to predict the three most critical performance indicators in sugar-beet agriculture: yield potential, α-amino nitrogen content, and polarisation (recoverable sugar content).
These models were developed using comprehensive training datasets collected across multiple growing seasons, incorporating thousands of field observations, satellite measurements, and laboratory analyses. The AI algorithms learn complex relationships between environmental conditions, crop development stages, and final quality outcomes—relationships that would be impossible for human observers to fully quantify.
The predictive capability of these models enables proactive farm management. Rather than reacting to problems after they occur, farmers can anticipate yield variations, identify areas requiring nutrient adjustment, and optimize harvest timing to maximize sugar recovery rates at the factory.
Validated Performance: Field-Proven Accuracy
The true measure of any agricultural AI system is its performance under real-world field conditions. Over two complete growing seasons at Kayseri Sugar's pilot sites, the SCI models demonstrated exceptional predictive accuracy that meets the demanding standards required for operational deployment in commercial agriculture.
0.85
Yield Model R²
Strong correlation with actual harvest results (MAPE: 15.3%)
0.81
α-amino N Model R²
Reliable nitrogen content prediction (MAPE: 18.9%)
0.88
Polarisation Model R²
Exceptional sugar recovery prediction (MAPE: 3.34%)
These R² values—ranging from 0.81 to 0.88—indicate that the models explain 81–88% of the variation in actual field measurements, representing field-deployable reliability. The low mean absolute percentage errors (MAPE) further confirm that predictions closely match observed values, giving farmers and factory managers confidence in using these forecasts for operational decision-making.
Dashboard Integration: From Data to Decisions
The sophisticated AI models developed through the SCI project have been seamlessly integrated into Agrovisio's web-based dashboard, creating an intuitive, user-friendly interface that translates complex data analytics into clear, actionable insights for agronomists and sugar factory managers.
Automatic Data Ingestion
System continuously collects satellite, sensor, and field data without manual intervention
AI Analysis Engine
Machine-learning models process incoming data to generate predictions and recommendations
Visual Map Interface
Field-level maps display yield zones, stress areas, and quality predictions with color coding
Report Generation
Automated creation of detailed reports for field-specific management decisions
This end-to-end automation dramatically reduces the time and expertise required to leverage advanced agricultural analytics. Field engineers can access the dashboard on tablets or smartphones, viewing real-time crop conditions and receiving specific recommendations for fertilizer application, irrigation scheduling, and harvest timing—all grounded in scientifically validated AI predictions.
Key Performance Indicators: What We Measure and Why
The SCI project focuses on three critical metrics that directly impact both farm profitability and factory efficiency. Understanding these indicators is essential to appreciating how the AI system creates value across the entire sugar-beet supply chain.
Yield prediction enables farmers to estimate total tonnage before harvest, supporting logistics planning and helping factories prepare processing capacity. α-amino nitrogen content affects sugar extraction efficiency—excessive nitrogen reduces sugar recovery and increases processing costs. Polarisation measures the actual recoverable sugar content, which directly determines payment to farmers and factory profitability.
By accurately predicting all three indicators several weeks before harvest, the SCI system allows both farmers and factories to optimize operations: adjusting harvest schedules, allocating processing capacity, and implementing late-season interventions when needed.
1
Crop Yield (t/ha)
Total biomass production determines harvest volume and logistics requirements
2
α-amino Nitrogen (mmol/100g)
Nitrogen excess reduces sugar extraction efficiency and increases factory processing costs
3
Polarisation (%)
Recoverable sugar content—the primary determinant of farmer payment and factory profitability
Building Digital Capacity Through Knowledge Transfer
Technology alone cannot transform agriculture—successful adoption requires building human capacity and fostering collaboration across the industry. The SCI project placed strong emphasis on training, dissemination, and stakeholder engagement to ensure that the developed tools would be widely understood and effectively utilized.
PANFEST 2024
International conference showcasing AI applications in sugar-beet production to research community
INFTEC 2024
Industry technology forum demonstrating practical implementation for factory managers and engineers
National Media Coverage
Public awareness campaign highlighting benefits of data-driven agriculture for sustainability
Beyond conferences and media outreach, the consortium conducted hands-on training sessions for field engineers at Kayseri Sugar Factory. These practical workshops equipped agronomists with the skills to interpret AI outputs, understand confidence levels in predictions, and translate model recommendations into specific fertilizer applications and irrigation schedules tailored to field conditions.
Toward Climate-Smart, Resource-Efficient Sugar Production
The SCI for Sustainable Sugar project demonstrates how advanced Earth observation and artificial intelligence can fundamentally transform traditional agricultural systems into climate-smart, resource-efficient production models that meet the challenges of 21st-century farming.
By providing precise, field-level insights into crop needs, the system enables targeted application of water and fertilizers—reducing waste, minimizing environmental impact, and lowering production costs. Farmers apply nutrients only where and when needed, rather than using blanket applications that often lead to over-fertilization in some areas and deficiencies in others.
This precision approach not only improves profitability but also addresses critical environmental concerns: reducing nitrogen runoff into water systems, minimizing greenhouse gas emissions from excess fertilizer production and application, and optimizing water use in regions facing increasing drought pressure due to climate change.
Resource Optimization
Targeted fertilizer and water application reduces input costs by 15–25% while maintaining yield.
Environmental Protection
Precision application minimizes nutrient runoff and reduces agricultural greenhouse gas emissions
Climate Resilience
Early stress detection and adaptive management help crops withstand increasingly variable weather
Economic Sustainability
Higher sugar recovery rates and lower costs strengthen farm profitability in competitive markets
The Future of Sugar-Beet Agriculture
The SCI for Sustainable Sugar project has successfully demonstrated that the integration of satellite monitoring, IoT sensors, laboratory analytics, and machine learning can create a reliable, field-deployable system for precision sugar-beet management. With validated R² values between 0.81 and 0.88, the models provide the accuracy needed for confident operational decision-making.
Looking ahead, the tools and methodologies developed through this consortium offer a scalable blueprint for modernizing sugar production across Europe and beyond. The dashboard interface can be adapted to different regions and growing conditions, while the underlying AI models can be continuously refined as more seasonal data becomes available. The collaboration between technology providers, academic researchers, and industry operators has created not just a software platform, but a sustainable framework for digital transformation in agriculture.
As climate variability increases and resource constraints intensify, data-driven precision agriculture will transition from competitive advantage to operational necessity. The SCI project demonstrates that this future is already achievable with existing technology—and that the benefits extend across the entire value chain, from field sustainability to factory efficiency to long-term industry viability.

For more information about the SCI for Sustainable Sugar project, including access to the dashboard platform, training materials, and research publications, contact the Agrovisio consortium or visit the project website.