Case Study: University of Lincoln achieves 95% AI model accuracy with V7 Darwin

How University of Lincoln Used V7 to Achieve 95% AI Model Accuracy

University of Lincoln’s LIAT team used V7 Darwin to tackle the difficult problem of predicting strawberry yields more accurately. Working in horticulture, they needed a better way to estimate fruit count, weight, maturity, and harvest timing from real-time crop imagery, since incorrect forecasts drive labor costs and waste.

Using V7’s labeling tools, including semi-automated video annotation, bounding boxes, polygons, directional vectors, and data versioning, the team built AI models for strawberry yield forecasting. With V7, they improved model accuracy from 85% to 95% and extended prediction time from 3 to 5 weeks, helping farmers make better decisions and reduce supply chain friction and crop waste.

University of Lincoln

Raymond Tunstill

PhD Researcher