Close to 12% of the world’s population struggles with hunger issues. While the number one cause of hunger can be attributed to poverty, food shortages also lead to people going hungry. The Food and Agricultural Organization of the United Nations has vowed to see that every human of the world has access to a nutritional diet. This is why agricultural sciences have become such a huge industry over the last decade. Simply planting food is not enough to feed the world, science is needed to boost crops. This past summer, the National Artificial Intelligence (AI) Research Institutes, a program funded by the National Science Foundation and the USDA’s National Institute of Food and Agriculture, and backed by the White House awarded a $20 million grant to the Center for Digital Agriculture at the University of Illinois to further develop a lab dedicated to ensuring agricultural stability.
The specific lab is called The Artificial Intelligence for Future Agricultural Resilience, Management, and Sustainability Institute (AIRFARMS), and is led by the Center for Digital Agriculture at U of I’s main campus, Urbana–Champaign (UIUC). UIUC is an R1 (Very High Research) ranked public university and the home of the fastest university based supercomputer. AIRFARMS “will serve as a nexus for multidisciplinary research teams that advance foundational AI and use these advances to address important challenges facing world agriculture,” as outlined on the lab’s website. Earlier this fall, a team of researchers from the program published their findings on using drones to create models that would help farmers know the best time to harvest crops.
Led by Assistant Professor Dr. Nicholas Martin, the crop the team focused on was soybeans. A highly important crop in the US and the world, soybeans grow easily, but tracking the plant’s maturity is very difficult. Often farmers wait until the end of the season to harvest their many plots of soybean because monitoring the corp is too time consuming. This leads to much of the crop being sub-par. Each plot, of which there could be thousands, needs to be checked twice a week to determine plant maturity. Specially trained individuals monitor the color of the soy pods to determine if they are ready to be harvested, a process that tends to have a wide margin of error.
Finding a way to streamline the monitoring of soybean plots would lead to more accurate planting and harvesting cycles. “Many research groups are trying to use drone pictures to assess maturity, but can’t do it at scale,” said Dr. Martin. “So we came up with a more precise way to do that. It was really cool, actually.” The first step was to take aerial images with a drone for 30 days at several growing sites. The images were then uploaded to a computer with an AI system that participating doctoral student Rodrigo Trevisan called deep convolutional neural networks, CNNs. As Rodrigo explains, this type of AI system tells a computer to recognize what makes up an image (color, texture, composition) the same way a human brain would.
“CNNs detect slight variations in color in addition to shapes, borders, and texture. For what we were trying to do, color was the most important thing,” Rodrigo said. “But the advantage of the artificial intelligence models we used is that it would be quite straightforward to use the same model to predict another trait, such as yield or lodging. So now that we have these models set up, it should be much easier for people to use the same architecture and the same strategy to do many more things.” The CNN model made an initial prediction on maturity based on the drone images taken on the first day. Then as more drone images were fed into the system, it was able to make an updated maturity prediction. In the end, using the drone and AI system the team was able to predict plant maturity within 2 days.
If the drones can aid in predicting the maturity of a crop such as soybeans, the implications it could have on predicting when to harvest plants like wheat are tremendous. This is exactly why the National Artificial Intelligence (AI) Research Institutes chose UIUC for their grant. At the Center for Digital Agriculture’s AIRFARMS lab, researchers will be coming up with solutions for world hunger issues. As UIUC Chancellor Robert J. Jones said, “Bridging the enormous potential of artificial intelligence with the science, engineering, and practice of agriculture offers the opportunity to solve some of the most critical challenges of our generation on a global scale. We’re very proud to be chosen by NSF as the home for this center that we truly believe will change the world’s agricultural future.”