Using hyperspectral imaging in winemaking

The winemaking process is an intricate balance of science and tradition. One of the most critical decisions a winemaker must make is determining the optimal time to harvest grapes. Traditionally, this involves manual sampling and chemical analysis, which are both time-consuming and provide limited data. However, a collaborative research project between the University of Southern Denmark (SDU), Capra Robotics, and our parent company Newtec Engineering A/S is transforming this process using advanced technology. At the heart of this innovation is Qtechnology’s hyperspectral camera, which is being integrated into a mobile robot platform to automate and optimize grape ripeness assessment.

Combining hyperspectral imaging technology and robots

The research team developed a robotic solution equipped with Qtechnology’s hyperspectral camera to scan vineyards and analyze grape ripeness. Unlike conventional cameras that capture three visible colors (red, green, and blue), hyperspectral cameras detect a broad range of wavelengths beyond human vision. This enables the technology to:

• Detect variations in sugar and acidity levels without physical contact.
• Provide a comprehensive, real-time analysis of grape maturity across the entire vineyard.
• Enhance decision-making by integrating the collected data into an AI-driven model for optimal harvest timing.

In early 2023, the team conducted initial field tests at Kimesbjerggaard Vineyard in South Funen, Denmark. The robotic system, fitted with Qtechnology’s hyperspectral camera, autonomously navigated the vineyard, capturing images of grape clusters. Researchers then cross-referenced these images with traditional chemical measurements to train the system’s AI model to recognize ripeness indicators accurately.

The impact: smarter, more efficient winemaking

The benefits of integrating hyperspectral imaging into viticulture are profound:

• Increased precision: Winemakers gain precise insights into grape ripeness across the entire vineyard.
• Optimized harvest timing: Ensures grapes are picked at their peak, improving yield and wine quality.
• Labor reduction: Automates the traditionally manual and resource-heavy process of grape analysis.
• Scalability: The technology can be adapted for use in other agricultural applications.

Future prospects

With further refinement, the robotic system aims to become fully autonomous, providing winemakers with real-time, AI-driven insights to optimize their harvest decisions. The success of this project demonstrates the potential for hyperspectral imaging to revolutionize not just viticulture but agriculture as a whole.

By integrating Qtechnology’s hyperspectral camera into a robotic vineyard monitoring system, SDU, Newtec, and Capra Robotics have introduced a groundbreaking approach to grape ripeness assessment. This collaboration underscores the power of technology to enhance agricultural efficiency, ensuring that the wine we enjoy is crafted with the highest level of precision and care.