How John Deere® is Using Machine Learning In Agriculture

Jerry Wallis
10 min read
Machine Learning Ag Spraying_crop1

Find out how equipment manufacturer, John Deere, uses machine learning in agriculture to develop innovative machinery and level up the precision spraying game.

How Is Machine Learning Used In Agriculture? 🧐

Today, machine learning is used in many industries to automate tasks and improve efficiency. Agriculture is one field that is particularly well-suited to machine learning, as it often involves repetitive tasks and large data sets. 

John Deere, one of the world’s leading agricultural equipment manufacturers, is using machine learning in agriculture to develop new machinery and optimise existing ones. For example, In this article, we’ll go over one of their innovations where they’ve used machine learning to create a more efficient crop-protection system that uses less herbicide and energy and delivers massive cost savings to the grower.  

In addition, John Deere is also using machine learning algorithms to automatically analyse sensors’ data to adjust agricultural equipment settings for optimal performance. 

For years, John Deere has been manufacturing innovative agricultural machinery to benefit their customers and the environment. They’ve done it again with The See & Spray™ Select and The See & Spray™ Ultimate technology systems. 

In this article, we’ll look at these two products in particular and conduct a deep dive into how the underlying technology used in these products is helping farmers produce better yields and improve profitability in the long run.   

Meet The See & Spray™ Select 🤝

Let’s first talk about the See & Spray Select, which uses a combination of computer vision and advanced technology to spray weeds in fallow fields. If you are unsure what this means, fallow fields are a common farming technique where the soil is simply left unplanted for a certain amount of time to restore organic matter and retain nutrients and moisture.

machine learning in agriculture 1
See & Spray™ Select

Essentially, in terms of what the spraying machine can do, the See & Spray™ Select system can detect weeds in the arable soil and exclusively spray just the identified weeds. The great thing about this select-spraying technology is that you only apply what you need where it is required, significantly reducing herbicide use.

machine learning in agriculture 2
See & Spray™ Select – Nozzle-Control Technology

Now, you might be thinking, how does the equipment know the difference between soil and weed? Don’t worry; we’ll now break down the  See and Spray Select™ workflow. 

The cameras are integrated into the boom scan every time the targeted spray is used. Data is sent continuously to the visual processing system, which detects the colour green and thus identifies the weeds in the fallow soil.

machine learning in agriculture 3
See & Spray™ Select Advanced Camera System

After weeds have been detected, a specific number of nozzles are used to target and spray them based on the concentration of weeds in an identified area. The in-cab display can be used to change the sprayer from selective to broadcast if there is a need to do so.

Of the area travelled, the in-cab display also records where the herbicide was applied and where it wasn’t. This information is streamed to the John Deere Operations Centre Account, allowing farmers to track performance, gauge spray efficiency, and make improvements based on decisions regarding weed control.

machine learning in agriculture 4
See & Spray™ Select In-Cab Display

All in all, a terrific way to manage weeds and improve farming performance, don’t you think? Just wait till you see what they’ve done with the following product.

The See & Spray™ Ultimate Took In-Crop Broadacre Spraying To A Whole New Level 🆙

A notch above the See & Spray™ Select is the See & Spray Ultimate. While these two products may sound very similar, they are pretty different in some crucial aspects. But we’ll get to that later; first, let’s understand what the See & Spray™ Ultimate is all about.

The See & Spray™ Ultimate uses technology designed by Blue River Technology, a subsidiary of John Deere. Like the See & Spray™ Select, a computer vision-aided camera system and processors are mounted onto the carbon fibre truss-style boom. 

However, the critical difference in the technology stack is the use of a machine learning system that detects weeds from crops and targets sprays, providing an effective weed-control strategy. Notice another difference? Yes, this equipment can detect weeds from crops, which is a level above the See & Spray™ Select.

machine learning in agriculture 5
See & Spray™ Ultimate

The visual processing system and camera setup are high-tech and designed for maximal efficiency. For instance, one camera is mounted every metre across the boom, which equates to 36 cameras on a 36.5-metre boom. This allows the system to scan upwards of 195 square metres at once. If that’s not being efficient, then we don’t know what is!

All this translates to a reduction in herbicide use which is essential for farmers looking to lower costs, reduce tendering stops, and cover more daily hectares. 

Furthermore, their nozzles use ExactApply technology, which reduces the chance of herbicide spilling over to unintended areas, thereby reducing the chances of collateral crop damage.

The ExactApply nozzle system was a spraying technology that John Deere introduced in 2017 to cover more of the target area with better precision to ensure every droplet does its job.

The See & Spray™ Ultimate continuously scans the field through camera vision technology to differentiate weeds from crops. Because of machine learning capabilities, it continually learns to be more accurate with each pass.

Now that we’ve been introduced to the See & Spray™ Ultimate let’s get a deeper understanding of how it works and uses computer vision and machine learning to protect crops from weeds effectively.

As the 36.5 metre-long boom moves across the field, 36 mounted cameras continuously capture high-resolution images, which are then sent to the graphics processing unit. 

machine learning in agriculture 6
See & Spray™ Ultimate’s Boom

Here, the machine learning model determines the presence or absence of weeds in the image based on prior knowledge and experience and signals target spraying if weeds are detected. All this happens at around 19 kilometres per hour, roughly equaling 185 square metres every second. Mind-boggling, right!

The machine learning model in the See & Spray™ Ultimate uses a technology John Deere likes to call Green-on-Green detection. Instead of just looking for green and spraying green, the machine learning model informs the cameras to specifically pick out the same weeds from earlier that need to be sprayed, thereby protecting the crops and yields. 

machine learning in agriculture 7
See & Spray™ Ultimate’s Dual Tank System

The See & Spray™ Ultimate also has a dual tank system that enables farmers to use two different tank mixes in a single pass. And lastly, using BoomTrac Ultimate, the See & Spray™ Ultimate provides better spraying accuracy than any comparable product because this technology lets the boom be kept within 250 millimetres of the target boom height resulting in precise spraying.

If you’d like to watch the See & Spray™ Ultimate in action, check out this fascinating video by John Deere.

While the See & Spray™ Select was already a smash-hit among farmers, John Deere went above and beyond to build an upgrade using contemporary technology that genuinely changes the agriculture industry and makes farmers’ lives much easier.

Note: As of now, the See & Spray™ Ultimate technology is yet to be launched in Australia, and it’s only available in the US for the time being. Hopefully, the company will make it available in the Australian marketplace in the near future.

Key Differences Between The See & Spray™ Select and See & Spray™ Ultimate 🤔

As we mentioned before, these are two pieces of machinery with some glaring similarities, such as the name or at least part of it, but they are very different from one another. To make it easier to understand, we’ve broken down these differences into the following categories.

Use-Case 📋

The See & Spray™ Ultimate is designed to detect weeds among only soybean, cotton, and corn plants, while the See & Spray™ Select is to be used in the fallow ground only, making it entirely unsuitable for in-season use.

Detection Process 🔬

The See & Spray™ Ultimate utilises cameras and processors to enhance accuracy, while See & Spray™ Select leverages colour-detecting technology to identify and target soil. 

Technology 👩‍💻

The See & Spray™ Ultimate also uses advanced Machine Learning technology, while the See & Spray™ Select relies on advanced nozzle control technology. 

Tank System 🚰

Last but not least, See & Spray™ Select has a single tank, while See & Spray™ Ultimate features a dual tank with different split-size options.

What Benefits Are Farmers Getting From These Products? 💰

Farmers have always had to be innovators, using the latest technology to increase yields and improve efficiency. Today, with equipment such as these, farmers can utilise sophisticated technological concepts such as machine learning to increase yields, protect crops, reduce pests and improve profitability. They can also leverage machine learning algorithms to analyse sensors’ data to identify crops’ problems and take corrective action. 

machine learning in agriculture 8
See & Spray™ Ultimate

But the benefits don’t end there, and here are a few more advantages farmers gain by incorporating innovative agriculture solutions such as the ones mentioned above in their workday.

Overall Benefits

Both products will help farmers reduce crop stress, improve yields, lower costs and cover more hectares daily.

Weed-Control 🌿

By eliminating weeds, crops can reap the full benefits of nutrients and moisture, allowing their roots to thrive and giving farmers the best possible yields.

Decreased Crop-Damage 🥦

With AutoTrac™ technologies of the See & Spray™ Ultimate, the sprayer stays in between rows. This means less chance of crop damage while making passes out in the field, which is a problem most farmers face when using other equipment.

Two Mixes In One Pass 💕

Farmers can simultaneously apply targeted spray and traditional broadcast with the dual tank configuration. 

For example, farmers can spray herbicide while broadcasting fungicide, combining two passes in one to save time and improve their productivity during each workday.

Real-Time Data Insights 📊

These products can stream agronomic data in the John Deere Operations Centre. This form of real-time analytics can help facilitate informed decision-making leading to improved performance down the line.

On top of that, the See & Spray™ Ultimate comes with a JDLink modem that can sync machine data to the Operations Centre for viewing and analysing purposes.

Significantly, this data can help farmers understand how much of the field is being covered in each pass and make appropriate changes to reduce overlap and maximise inputs every workday.

Conclusion 🙋‍♂️

In an industry that may often be unfairly overlooked regarding technological innovation, John Deere has consistently pulled off the unthinkable and is providing farmers with high-quality equipment designed to maximise yields and minimise costs.

Companies like John Deere are a testament to how innovation and technology can be used to achieve excellence in any industry. As advanced technology such as machine learning continues to develop, innovative companies like John Deere will play an increasingly important role in the future.

If you’d like to learn more about machine learning and technological innovation and how it can help businesses like yours scale and inspire, then reach out to us. We’ll be happy to chat over a discovery session and get you a step closer to building innovative customer experiences.

Please note: All images and registered trademarks are the copyright and ownership of John Deere®. They are used in this article entirely for the purpose of education and review of the technology being utilised by the company in its products.

Published On

July 25, 2022