machinery for use in the agricultural, construction and forestry industries. ... By monitoring machines remotely as they
IN ASSOCIATION WITH:
MACHINE LEARNING IN AGRICULTURE INTRODUCTION John May is the president of agricultural solutions and chief information officer of John Deere, the Illinois-based global manufacturer of machinery for use in the agricultural, construction and forestry industries. Among other responsibilities, he oversees the company’s precision agriculture technology and cloud platform, which enables customers to collect, analyze and share data in real time.
WHAT CHALLENGES LIE AHEAD IN AGRICULTURE? Looking out to 2050, the demand for food could double if we grow to 9 billion-plus people on the planet. We could also see better diets around the world as societies improve from an economic standpoint. There will be tremendous pressure to produce more food than we’re producing today. We believe that precision agriculture will be one of the key technologies to enable that increased food production. We’ll go from managing a field, or maybe today a section of a field, to managing every individual plant to maximize its yield. Precision agriculture is a critical strategy for farmers to unlock value and productivity in a sustainable way.
HOW DOES MACHINE LEARNING FIT INTO JOHN DEERE’S PRECISION AGRICULTURE STRATEGY? Farming is an annual planning process based on multiple years of data. Throughout the year farmers are analyzing that data and asking themselves: What seed should I use? How should I space that seed? What depth should I [plant at] to maximize the yield? Today farmers rely on a variety of trusted advisors such as agronomists to help them make the right decisions. AI will allow farmers to sense all the variables on a real-time basis and make decisions on the spot as they go through the field thanks to a smart combine or sprayer. Farmers and their trusted advisors will get data even faster. The more technology advances, and the more the data can become real time, the better decisions you can make.
WHAT OTHER TECHNOLOGIES ARE AT PLAY IN THIS MACHINE LEARNING ECOSYSTEM?
JOHN MAY PRESIDENT OF AGRICULTURAL SOLUTIONS AND CHIEF INFORMATION OFFICER, JOHN DEERE
For the last 20 years, we’ve been focused on building out a suite of technologies. One is autonomous driving capability. Today our vehicles are guiding themselves through the field. This gives the farmer precision because, path after path after path, the vehicle will follow the same route. Geospatial reference enables farmers to know exactly what they’ve done in a particular area. Another technology
advancement is onboard computing. Tremendous amounts of data are produced through a whole suite of sensors in each pass through the field; for example, now you can know where every individual seed is in the ground. Another core technology is telematics, the power to send information — location and usage tracking, fleet driver management, machine performance and diagnostics — to the cloud or to a mobile device, wherever you want to send the information. Telematics unlocks that. A farmer in a combine, an agronomist studying the crop or a dealer can all access the data to work more efficiently and solve problems before they arise.
IS ALL THE DATA FEEDING PREDICTIVE ANALYTICS? By monitoring machines remotely as they’re in the field, we can maximize their functions. We also have technology that allows farmers and our dealers to remotely access a machine — a combine, for example — to see what’s happening. That combine is linking back to the cloud to optimize itself. It’s giving the customer recommendations, such as ideal settings for harvesting, as it’s going through the field. We can do a lot of predictive analysis and sense mechanical failures before they even happen. Being able to monitor how the machines are performing is a huge benefit to us as a company and also a big benefit to our customers.
TELL US MORE ABOUT THE EFFICIENCY COMPONENT THAT MACHINE LEARNING BRINGS TO THE FARM. AI fits in with what we call job optimization. There are a lot of variables coming at farmers, whether they’re planting, spraying, combining or tilling. Each step in the production process can be difficult. To optimize that job, you have to look at different variables — for example, harvesting. A combine takes the grain out of the field, but it has to work in an environment that has a lot of variability and challenges. So the question is how to automate this and make it more precise, easier and smarter to ensure a machine works the best that it can, doing the job it is intended to do.
HOW IS THE VALUE THAT COMES FROM DATA COLLECTION AND ANALYSIS UNLOCKED BY THE CLOUD? We’re building out an ecosystem that seamlessly connects equipment, technology, people and insight. The John Deere
Operations Center is at the heart of that effort. It allows farmers to send data from the machines directly to the cloud. Now the farmer can share that data with any trusted advisor — an agronomist, for example — to make better decisions on whatever production step they’re in and improve the outcome. The Operations Center is about seeing data, supporting collaboration in an open platform and providing farmers a way in which to direct data where they want and need it all, so they can gain insight and then execute upon it in an increasingly automated fashion.
WHAT IS THE ROLE OF COMPUTER VISION IN YOUR PRECISION AGRICULTURE STRATEGY? Machine learning capabilities in combination with computer vision can enhance the required in-field tasks, such as planting, protecting and harvesting. We will have computer vision technology on all of our equipment in the future to unlock even more value for our customers through automation. Computer vision, robotics and machine learning will offer an advanced set of eyes and knowledge that will optimize farm machinery. John Deere recently acquired Blue River Technology and its See & Spray technology, which can differentiate between a weed and a plant. On a real-time basis, using the data you’re capturing in milliseconds with computer vision, the spray will target just the weeds and dramatically reduce the amount of herbicides used in agriculture. Automation will take productivity to a level we haven’t seen in agriculture, one where you’re making sure that a plant — whether soybean, corn, cotton or any other crop — has an environment that allows it to grow to its fullest potential.
WHICH HAS SIGNIFICANT FINANCIAL BENEFITS. WHAT IMPACT IS MACHINE LEARNING-DRIVEN PRECISION AGRICULTURE HAVING ON BALANCE SHEETS? Blue River has been working on optimizing a sprayer using artificial intelligence. Based on research we’ve done, and based on what Blue River has proven in the field, our customers have the opportunity to eliminate up to 90% of the cost of what they were spending with prior technology in order to spray a field with herbicide, for example. They’re only putting down 10% of the herbicide they used to in the past. It has a big benefit to the yield of the crop. From a revenue and cost standpoint, you’re maximizing yield by creating that environment where you grow the healthiest plant. So it’s a huge savings to our customers.
