Yield Prediction for all Mankind: Part I

The total agriculture land in the US is 1 billion acres, of this 600 million is for grazing and 400 million is for crop production. Variations in the yield of this area by just 1% can have huge impacts on food supply both domestically and globally. Predicting the yield can help reduce the impact.

Yield predictions on crops are nothing new. They have been performed on numerous crops such as corn, wheat, and rice for many years now. This is due to many reasons such as versatility of these crops and government programs. If yields can accurately be predicted, there are many applications beyond food security: insurance companies can reduce risk in farms insured, harvest logistics/infrastructure can be optimized and there can be a viable crop commodity trade market.

The problem is status quo.

Most yield predictions are made manually by counting clusters, sections, trees, etc. and extrapolating out to the whole field. This method results in high and unpredictable errors. As in most cases, this is considered ‘high-tech.’ Agriculture has been slower than other industries to adopt new technology. Case in point, GPS guided tractors are just now becoming mainstream. GPS units in commercial vehicles have been around for 10+ years. The Ag industry is very much a tradition as is it a way of life and most traditions don’t change. It has been done a certain way for generations. However, with more demand, emerging global countries, drought, war, you name it, more is needed to feed the world.

At Vinsight, we believe that knowing how much is grown and optimizing the output will help us in securing the world’s food supply. In order to get there, better yield prediction is necessary.