Predictive Ai: harvests & prices

In the Predictive Ai: harvest & prices project, the partners are developing computer vision-based products and solutions for the greenhouse sector to enable operational optimisation. The developments resulting from the project enable greenhouse owners to automate the counting and measuring of fruit and vegetables using mobile cameras – two of the most time-consuming and complicated tasks currently involved in greenhouse cultivation. An extensive client discovery process in the Netherlands and Germany revealed that there is a high demand for automation. For the prediction models, computer vision (CV) is combined with artificial intelligence (AI) to predict and even influence the growth rates of fruit and vegetables. Mobile components such as special camera drones, camera towers and camera robots are added to make the system cost-effective. An Android app for manual data entry complements the mobile components. The project partners use CV to understand the growth rate of fruit and vegetables in greenhouses so that greenhouse owners can predict and even influence when they can harvest how much by adjusting the appropriate inputs (e.g. light). This addresses important issues in the greenhouse industry, such as reducing the currently significant product losses due to discrepancies between harvest and sale times, and enabling greenhouse operators to make more accurate commitments to buyers, such as supermarket chains, about when the greenhouse will deliver how many products. Furthermore, the transferability of image processing to different crops and a link to price predictions is being developed to better predict and plan revenues.