HarvestAi, a German deep-tech startup focusing on AI solutions for indoor farming operations, and Gemüseproduktion Zorbau, a member of the Gemüsering Group, an internationally acting vegetable and fruit producer and distributor, jointly develop AI-powered forecasting technology for a commercial tomato greenhouse in Germany. Since September 2022, these two innovative entities have been working together in stealth mode and made great progress in achieving accurate crop growth and yield prediction.

Belonging to the Gemüsering group, Gemüseproduktion Zorbau has been at the forefront of integrating cutting-edge technologies into its farming operations to ensure efficient and innovative solutions. Their collaboration with HarvestAi, a leader in developing advanced machine learning and computer vision-powered crop growth and yield prediction models, marks a significant leap forward in indoor farming technology.

HarvestAi’s Crop Growth and Yield Prediction technology leverages a wealth of data from Gemüseproduktion Zorbau’s greenhouse, including harvesting, climate, and crop registration data. The unique aspect of this collaboration is the autonomous gathering of crop registration data through state-of-the-art computer vision, utilizing cameras installed in the greenhouse. Additionally, APIs have been implemented to collect harvesting and climate data, while external weather data is also incorporated, allowing the consideration of changing light and temperature conditions that affect greenhouse operations.

Dr. Georg Caspary, CEO and founder of HarvestAi, shares his enthusiasm: “This partnership is a testament to our commitment to revolutionize indoor farming using AI. The integration of our technology with Gemüseproduktion Zorbau’s operations opens new horizons in precision indoor agriculture, enabling smarter, data-driven decisions that were not possible before.”

After months of meticulous data gathering and joint user workshops, HarvestAi and Gemüseproduktion Zorbau have made remarkable progress. The precision of their forecasting model, predicting when and how much tomatoes of a specific quality can be harvested, now exceeds 90%, using machine learning only. Integrating Computer Vision has pushed this accuracy even further. Their technology not only forecasts immediate harvests but also identifies longer-term trends. This breakthrough allows for data-informed decision-making, enhancing communication between the growth and sales teams at Gemüseproduktion Zorbau and enabling them to engage in negotiations with customers earlier. The application of computer vision also allows real-time insights into the availability of ripe and unripe fruit in the greenhouse. With a size of more than 8 hectares, this allows enormous efficiency gains and frees up more capacity for the growth team at Gemüseproduktion Zorbau.

Furthermore, the system offers simulation capabilities for growers setting greenhouse climate conditions in the HarvestAi interface and instantaneously seeing the resulting impact on plant growth and harvest events. This not only aligns growth strategies with market requirements but also enables growers to test various growth scenarios for cost-effective production, especially during periods of high energy costs and unfavorable external weather conditions that affect the light intensity and temperature within the greenhouse.

Dr. Lukas Scholz, CEO of Gemüseproduktion Zorbau, states, “Our collaboration with HarvestAi is a game-changer for our operations. The insights and efficiencies gained through this technology are invaluable, empowering us to meet market demands more effectively while optimizing our resource usage.”