How Artificial Intelligence Can Help Feed The World

How can the use of artificial intelligence (AI) ensure global food security? AI is good at processing vast quantities of information to see patterns. As such given environmental variables that expose an AI to all that nature can throw at farmers to negatively impact crop yields, the technology can be used to spot crop diseases, infestations, contamination risks and more and initiate or suggest preemptive action.

What tools can produce the data that makes it possible for AI to help with food security? AI can use computer vision to review satellite images, and real-time video from drones. AI can review information fed to it in real time from in-the-field sensors. Machine learning can compare what it learns from legacy datasets with current field observations to help it spot plant stress caused by various issues such as lack of water, poor soil conditions, pest infestations, and negative weather conditions.

Fermata is an Israel-based company that uses computer vision and AI to monitor crops grown in greenhouses. Croptimus, its adaptive computer vision platform, automatically detects pests and diseases at early stages and informs farmers to take action. To date, users of Croptimus have reduced leafy green tomato, and medicinal plant crop losses by 30%.

Valeria Kogan, founder and CEO of Fermata, was approached in 2017 by a tomato grower friend who asked if Kogan could invent a program to help keep plants healthy, identify risks, and reduce crop losses. Kogan looked at the technical challenge and what tools were available and developed bioinformatic AI algorithms that could be married to off-the-shelf vision technology for monitoring crops. She began working with greenhouse growers and considered deploying a field robot for plant surveillance. This proved to be overkill. No one needed a robot for field surveillance when a camera system capable of a 360-degree view could do the same.

The problem of using the camera data feed to identify pest infestations and plant diseases represented the bigger challenge. What was needed was a trainable AI algorithm and lots of publicly available data to feed it. That way the AI could use machine learning to identify the range of threats. Human input was needed to label the individual threats before the software could begin to do its job. After Fermata launched in 2019 a research centre containing infected plants was added for the AI to continue to learn over over and above what it observed from collected field data.

The key to Croptimus is the ability to identify lots of problems through gathered data. Kogan notes that up to 40% of what farmers grow succumbs to pests, diseases and other environmental stresses. These are preventable losses. For the company to identify them all it required talking to growers. She describes what this was like in a recent article published in TechCrunch where she noted:

“We love customers who have a lot of problems because it brings us a lot of data, especially if they have some deadly diseases…We’re like, okay, let’s look serious on the call, and that we’re worried because that is bad news for them but for us it’s amazing news.”

Croptimus provides 24/7 field observation of crops. Its precision 2D mapping combines visualization, data collection, and analysis to identify pest and disease incidents. Data is continuously filtered through its database in the Cloud where it is analyzed. Problems are identified by type and precise location. Notifications are sent immediately to the grower. After intervention and mitigation, Croptimus continues to monitor the effectiveness of the intervention and ensure there is no recurrence.

The list of crop issues that Croptimus has identified so far includes: Aphids, Spider Mites, Thrips, White Fly, Powdery Mildew, Botrytis, and Mosaic. Croptimus can also identify nutrient issues and mechanical damage to crops. The company is working with AI to add Hop Latent Viroid to the database.

This work is one case example of how AI is being applied to precision agriculture as farmers seek help in dealing with the challenges of providing food security while facing environmental changes exacerbated by climate change in the 21st century.