Software to manage the robot’s data, predict yield, and teach the robot to recognize new traits.
TERRA-MEPP is developing high-throughput machine learning software to analyze robot data to predict crop performance and yield. The project is also developing BioCro, an R package for simulating and optimizing photosynthesis (leaf-level and canopy) to accurately predict the growth of biomass crops. The model has predicted yield using the first seven weeks of data and the first twelve weeks of data, with a Spearman's rank correlation for prediction of the end-of-season yield of .56 and .86 respectively. License our biophysical sorghum model to predict yield throughout the season and more accurately project yield futures.