CES in Las Vegas is the world's largest tech show, with awesome displays of artificial intelligence, machine learning, and autonomous vehicles. This year, my favorite example of all these futuristic technologies comes from a company many associate more with Green Acres than Silicon Valley.
John Deere (DE +0.88%) may be 186 years old, but it wowed CES visitors with its massive combine tractor that saves 60% in chemical costs by using AI, NVIDIA processors, and sensors to spray chemicals with pinpoint control. The company also showed off other futuristic technology, including its self-driving tractor and an app that allows farmers to monitor all aspects of their tractors in the fields.
Watch the video below for a guided technological tour from Julian Sanchez, director of emerging technologies in John Deere's Intelligent Solutions Group (ISG).
A full transcript follows the video.
Julian Sanchez:
[00:00:00] We're going to get real close and personal to the technology that we call See & Spray. See & Spray is a technology that gets mounted to a sprayer, sprayers used to drive through a field and apply herbicides or nutrients to plants. What we've done is we've added 36 cameras to this 120 foot carbon fiber boom, and we've added 10 NVIDIA GPUs that are housed in a special case and they keeps them passively cooled so that they can withstand very harsh conditions. As this machine is moving through the field at 12 miles an hour, those cameras are looking ahead, and they're looking to detect weeds on the ground. If you look over there, those x's represent weeds. The machine detects those weeds and applies herbicides only to the weeds. The herbicides [00:01:00] come out from underneath here in these nozzles. Every time the camera sees a weed, it sends a message, and within a millisecond it activates these nozzles and then sprays only the weeds.
This system has the ability and potential to save farmers 2/3 of the herbicides their use. Huge economic impact, but even more of an environmental impact. The magic in developing the system is we had to spend several years capturing images of what different weeds look like all over the world, and training a deep neural network in this is what weeds look like, this is what a plant looks like. Again, here you see a visual representation. Let's assume this machine was moving at 12 miles an hour. That's the speed at which we could operate using the system. The rows, those represent crops, those are plants, could be corn, could be soybeans. Every once in the middle there you see weeds.
[00:02:00] Again, that's what those cameras are looking for and they're only spraying those weeds. Again, deep neural network, one of the most sophisticated artificial intelligence system that is real is being used by farmers already today. It works extremely well in the harshest of conditions. Again, we're talking about dusty fields, crosswinds, different lighting conditions.
The ruggedization of this advanced AI system on the edge is tremendous. One last important point is all of the vehicles that we make come with connectivity. All of this data does get transmitted to the Cloud so the farmer can review it later. However, because we're in rural environments, we can't rely on connectivity all the time. All the compute is done at the edge.
This system is also intended to save on chemicals. What happens when this is a planter and [00:03:00] as this moves through the field, it's dropping seeds, it drops 30 seeds per second. Traditionally, what happens is, something that called starter fertilizer gets applied. Can we move around you real quick here just to see? Can you run it? The seeds are going to be coming out at 30 per second. Traditionally, what happen is you run that. Is there a thing? [inaudible 00:03:30] second. Got you. Yeah, perfect. Right now, the technology, as of today, we just continuously apply starter fertilizer. You'll see it in the second, it'll start to pulse. Now we time each seed with a pulse. Again, this also has the opportunity to save 60 percent and starter fertilizer.
Every sensor on these machines is connected [00:04:00] to a mobile application. This is a representation of what a farmer will be seen on their phone. At any point, a farmer can basically from anywhere in the world monitor what's going on with any of their vehicles. They can click on any one of those vehicles. They can monitor the data that's coming from any of those sensors. In some cases, they can reconfigure the machine remotely from the mobile app. If you asked 10 farmers today, what's the coolest piece of technology and agriculture? I bet at least half of them would tell you it's the mobile app. Because the mobile app brings it all together.
Rex Moore: How about autonomous tractors?
Julian Sanchez: Let's go over here. Let's talk about autonomy then. Last year at CES, we announced our first full autonomous tractor. This is a sensor suite for the autonomous tractor. This sensor suite will go in the front of the tractor, and it has three pairs of stereo cameras. There's actually another [00:05:00] set of stereo cameras in the back, so we get a 360 view. But basically, we've introduced autonomy in one of the applications in agriculture, which is Tillage. Over the next decade we plan to extend autonomy to the other applications in agriculture which is spraying, harvesting the crops and planting the crops. Just about every farmer you talk to globally, one of the top three asks that come up is autonomy. Because labor shortages are an issue. By the way, labor shortages in agriculture were an issue before they were labor shortages everywhere else.
