It is predicted that the precision agriculture market will reach $12.9 billion by 2027. With this increase, there is a need for sophisticated data-analysis solutions that are capable of guiding management decisions in real-time. A new methodology has been developed by an interdisciplinary group at the University of Illinois, and it aims to efficiently and accurately process precision agricultural data.
Nicolas Martin is an assistant professor in the Department of Crop Sciences at Illinois and co-author of the study.
“We’re trying to change how people run agronomic research. Instead of establishing a small field plot, running statistics, and publishing the means, what we’re trying to do involves the farmer far more directly. We are running experiments with farmers’ machinery in their own fields. We can detect site-specific responses to different inputs. And we can see whether there’s a response in different parts of the field,” he says.
“We developed methodology using deep learning to generate yield predictions. It incorporates information from different topographic variables, soil electroconductivity, as well as nitrogen and seed rate treatments we applied throughout nine Midwestern corn fields.”
The team used 2017 and 2018 data from the Data Intensive Farm Management project to help develop their approach. In that project, seeds and nitrogen fertilizer were applied at different rates across 226 fields. Those fields were in different areas of the world, including the Midwest, Brazil, Argentina, and South Africa. High-resolution satellite images were provided by PlanetLab, and they were paired with on-ground measurements in order to predict yield.
The fields were digitally separated into 5-meter squares. The computer was given data on soil, elevation, nitrogen application rate, and seed rate for each square, and it then began to learn how the yield in that square is determined by the interaction of the factors.
In order to complete their analysis, the researchers relied on a convolutional neural network (CNN). A CNN is a type of machine learning or artificial intelligence. While some types of machine learning get computers to add new data into existing patterns, convolutional neural networks do not take existing patterns into account. CNN’s look at data and learn the patterns that are responsible for organizing it, and it works in a similar way to how humans organize information through neural networks within the brain. The CNN approach was able to predict yield with a high accuracy rate, and it was compared to other machine learning algorithms and traditional statistical techniques.
“We don’t really know what is causing differences in yield responses to inputs across a field. Sometimes people have an idea that a certain spot should respond really strongly to nitrogen and it doesn’t, or vice versa. The CNN can pick up on hidden patterns that may be causing a response,” Martin says. “And when we compared several methods, we found out that the CNN was working very well to explain yield variation.”
The use of artificial intelligence to analyze data from precision agriculture is a new field, but it is one that is growing. Agriculture is one of the major industries that will be drastically changed by artificial intelligence, and the use of it is continuing to increase. According to Martin, this experiment is just the start of CNN’s being used in a variety of different applications.
“Eventually, we could use it to come up with optimum recommendations for a given combination of inputs and site constraints.”