The new AI algorithm predicts corn yields better, Because some reports estimate that the precision agriculture market will reach $ 12.9 billion by 2027, there is an increasing need to develop sophisticated data analysis solutions to control real time management decisions.
A new study by the interdisciplinary research team at the University of Illinois offers a promising approach to accurate and efficient Ag data processing.
Researchers say: We are trying to change the way people conduct agronomic research. Instead of making a small field, compiling statistics and publishing funds, we try to involve farmers more directly.
We conduct experiments with agricultural machines in their own fields. We can find location-specific answers for various inputs. And we can see whether there are reactions in various parts of the field. The new AI algorithm predicts corn yields better.
The researcher adds: We have developed a methodology that uses in-depth training to produce estimated earnings.
It contains information on various topographic variables, soil electrical conductivity, and nitrogen and seed treatments that we have implemented in nine cornfields in the Midwest.
The research team is working with data for 2017 and 2018 from the Intensive Data Agriculture Management project, where seed and nitrogen fertilizer are spread at different speeds in 226 fields in the Midwest, Brazil, Argentina and South Africa.
Terrestrial measurements are paired with high-resolution satellite imagery from PlanetLab to predict results.
The fields are digitally divided into 5 square meters (about 16 feet). Soil, altitude, nitrogen and seed data are sent to the computer for each box to study how factors interact to predict the results in that box.
The researchers approached their analysis with a type of machine learning or artificial intelligence known as the Convolutionary Neural Network (CNN).
Some types of machine learning begin with a model and ask the computer to adapt new data bits to this existing model. Convolutional neural networks are blind to existing models. Instead, they take data and learn the patterns they set, just as people organize new information about neural networks in the brain. The new AI algorithm predicts corn yields better.
The CNN process, which predicts high precision mining, has also been compared with other machine learning algorithms and traditional statistical techniques.
In fact, we don’t know what causes the difference in input data mining responses in the field.
Sometimes people have the idea that certain locations must react strongly to nitrogen and not – or vice versa.
CNN can find hidden patterns that can cause reactions, say researchers. Comparing several methods, we found that CNN works very well to explain the differences in results.
Using artificial intelligence to reveal precision agriculture data is still relatively new, but Martin said that his experiments were only the tip of the iceberg in terms of potential CNN applications.
Finally, we can use it to make optimal recommendations for a combination of site entries and restrictions.