3DgrainsML is a technology that combines artificial intelligence and high-resolution 3D scanning to study the shape of cereal grains. By analyzing grain shape, the project can determine the environmental and farming conditions under which crops were grown. This provides a non-destructive and low-cost way to study agriculture, climate impacts and food production strategies, using grains that are widely available today and in historical records.

Problem

Understanding how crops react to growing conditions such as water availability, soil management and climate is essential for agriculture, sustainability and food quality. However, many existing phenotyping methods are expensive, destructive or difficult to apply at scale. This limits their use in agronomy, food provenance, quality control and long-term studies of past agricultural and climate conditions.

Solution

3DgrainsML combines high-resolution 3D scanning of grains and AI to classify grains according to their growing conditions. Each grain is scanned and digitally aligned so shapes can be compared consistently. The system then translates grain shape into numerical features using mathematical techniques that capture even small differences in form. Machine learning models use these features to classify grains according to growing conditions, such as water use, fertilization practices and climate-related factors. Because the approach does not harm samples, it may be used on both present and historical grain collections, enabling the study of previous agriculture and climate.