Improving crop insights and yields with Artificial Intelligence

Farms across the United States and worldwide have been using advanced technologies to help improve crop yields, maximize resource efficiency, and more. The National Institute of Food and Agriculture lists robotics, GPS-related solutions, and IoT capabilities as some of the technologies that can increase crop productivity while reducing the amount of fertilizer and water used for more environmentally sustainable operations and increased supply of food for consumers.

“Growing up on a farm in Nebraska and being interested in technology, I’ve seen that intelligent solutions can have a huge impact on agricultural performance,” says Jeremy Groeteke, U.S. Digital Ag Lead at Corteva Agriscience. “When my dad started out, he had a four-row planter and six-row combine. Today he has multiple 24-row, self-driving planters, automated precision applications of nutrients in the crops, and more, and doubled his yield.”

Corteva Agriscience, a global pure-play agriculture company, has been working with farmers to establish more useful and powerful agriculture technologies. It has used Google Cloud solutions across many areas of its company, including in R&D, data science, and the plant breeding department. Jeremy and his department utilized the tools from Google Cloud and partner Kin + Carta for a project that would use AI to provide corn farmers with a more advanced approach to yield management and predictions.

Before reaching out to Kin + Carta for its specializations in customer service, user research, and application development, Corteva engaged the tools from Google for some of the backend technology that went into the app such as machine learning solutions, crash analytics, computer vision, and more.

The collaboration provided the opportunity to leverage Kin + Carta’s leading approach to user experience, the tools from Google machine learning technologies and Corteva’s skill in data science and agriculture. Corteva views the Pioneer Seeds app as being best-in-class when it comes to usefulness and usability among the farming community.

Firebase platform gives Corteva confidence that its app is performing properly, as well as visibility into feature adoption that informs ongoing improvements. Firebase Crashlytics provides real-time performance monitoring that makes any outages or other issues known in a moment’s notice.

“We knew early on that we would need to be able to deliver the app’s functionality to farmers who don’t have internet connectivity, as that can be challenging in rural areas,” says Jeremy.

Working with Kin + Carta and TensorFlow Lite, Corteva developed and deployed algorithms that can count the kernels and provide results even when a farmer is disconnected.

“Now, a grower can literally snap a picture of an ear of corn, and the app, along with our algorithms, combines the information from that image with other data, including satellite imagery, to provide results instantly,” says Jeremy. “All our growers, regardless of which device they’re using, receive the instant value that we needed to provide to make this app a success.”

With more than 20,000 users, Pioneer Seeds is now the largest app that Corteva has launched. Farmers have estimated over 121,000 different crop yields more accurately and further into the future than before, with farmers gaining insights up to two months before harvest that can also indicate when different strategies need to be taken to improve output.

“The Corn Yield Estimator feature in the Pioneer Seeds app launched, bringing the machine-learning model developed by our R&D team to farmers to improve the speed and accuracy of yield estimates in the field,” says John Arbuckle, Vice President of Field Technology Innovations and Operations at Corteva Agriscience.

The Pioneer Seeds app is one of the many ways in which Corteva is working to improve the outlook for farmers and provides proof of how quickly innovation can make a difference when it is done right.

“Agriculture is sometimes viewed as lagging behind in technology adoption, but this could not be further from the truth,” says Jeremy. “Farmers are great adopters of technology and were even some of the first users of self-driving technology in the form of AI-powered combines. The answer is to offer technology that provides instant and measurable value. If it does, they will adopt it.”

Corteva is now working to further improve and expand the features of the app through more advanced satellite imagery and solutions. It hopes to provide even more useful information to farmers about any issues a crop might be encountering and ways to solve those problems. Jeremy believes that better communicating the value of a given technology will help to improve adoption and, ultimately, agricultural performance.

“It is a key role for our organization to identify and create technologies to put into our growers’ hands that make them more efficient and profitable,” says Jeremy. “With technologies like drones, IoT, AI, financial analytics, and more, this decade is going to see even more breakthroughs in agriculture.”

• Enables yield predictions two months before harvest
• Automates kernel counting using computer vision algorithms built with TensorFlow for faster, more efficient crop intelligence
• Provides best-in-class application to farmers without the need for internet connectivity
• Powers reliable app with real-time monitoring capabilities through Firebase platform
• Combines best-in-class cloud compute and ML with the power of the intelligent edge to deliver first-in-class optimization for agronomists