“With AI technology, anyone can be a programmer. If you have an idea, whether you can code or not, AI can write the RealSense code, and you can deploy an AI-enabled 3D camera for less than $100.”

Walk any robotics show floor and chances are you will see RealSense cameras in many of the robots and autonomous mobile robots (AMRs) on display, which made it surprising years ago when erroneous reports said that Intel was winding down the RealSense division. Things have changed drastically since then: RealSense just recently spun out of Intel as a standalone business, with no plans of slowing down.

In this episode of the Manufacturing Matters podcast, Chris Matthieu, chief developer evangelist, and Fred Angelopoulos, vice president of global sales, join TECH B2B Marketing’s Winn Hardin and Jimmy Carroll to discuss the future of RealSense, along with topics including the state of 3D imaging, the convergence of AI and 3D technology, and recent innovative applications that leverage RealSense technology. Additionally, the discussion covers humanoid developments, advice for startup companies, digital twins, and more.

Jimmy Carroll: [00:00:06] Hi everybody. My name is Jimmy Carroll. I’m the vice president of operations at Tech B2B Marketing. And welcome to this episode of the Manufacturing Matters podcast, where I have the pleasure of being joined by my friend and co-host Winn Hardin and Chris Matthieu and Fred Angelopoulos from Intel RealSense. Guys, thanks so much for taking the time. I really appreciate it. So tell us a little bit about what you’re most excited about right now at Intel RealSense and what’s new and fun.

 

Chris Matthieu: [00:00:33] Well, I’m excited about tech. I’m a hacker on the team, chief developer evangelist, and we’ve got a bunch of new technology that we just released at the show. We launched a Power over Ethernet camera. We’ve got some industrial cameras, we’ve got a facial authentication camera, a lot of new cool toys, very powerful toys. For me they’re toys. For everyone else they’re really industrial machines and robot sensors.

 

Winn Hardin: [00:01:00] Are you guys building those completely in-house at Intel, or are you guys partnering in terms of the camera builds?

 

Chris Matthieu: [00:01:05] We build it ourselves. We have manufacturing facilities across the world.

 

Winn Hardin: [00:01:09] That I know. So what are you guys most excited about here at Automate this year so far?

 

Chris Matthieu: [00:01:19] I can tell you we walked the show floor, and almost every single robot or arm or AMR had RealSense cameras in them. So we were doing sound bites. Everyone was like, “We love RealSense.” I’m like, “Oh, fantastic. A-plus. You get a gold star. You get a gold star.”

 

Winn Hardin: [00:01:35] Everyone has been waiting for volume availability for a long, it seems like a long time. Although I’m just old.

 

Jimmy Carroll: [00:01:43] It’s almost like — and this is maybe like waxing poetic a bit — but with the advent of the Intel RealSense it was almost like 3D imaging sort of came of age, like affordability became an option for 3D and a smaller device, not a larger, more expensive device. And it’s just opened up a lot of doors, I think, for a lot of people.

 

Chris Matthieu: [00:02:04] You know, earlier this year we launched, it’s called a D421 camera, $80, legit RealSense stereo depth camera. I mean, all the makers, hackers, experimenting students they all gravitate to that camera. And it’s every bit as powerful as all the rest of our products.

 

Winn Hardin: [00:02:23] Is that a camera onboard or is that a full enclosure?

 

Chris Matthieu: [00:02:25] It’s open, but it’s got a metal plate on the back with mounting holes. So it’s perfect for Raspberry Pis or small AMRs, experiments, and whatever you build with that camera on a smaller $200 robot, we’ll just apply to a $50,000 AMR. As long as it’s runing ROS, in real sense it doesn’t matter the scale.

 

Winn Hardin: [00:02:49] Beautiful. That’s so cool.

 

Jimmy Carroll: [00:02:51] Yeah, it’s like the democratization of 3D imaging, right?

 

Chris Matthieu: [00:02:54] A lot of people think of robotics and they think of humanoids, which of course is the coolest, but there’s a lot of small robotic applications.

 

Winn Hardin: [00:03:05] The vast majority of them.

 

Fred Angelopoulos: [00:03:07] Yeah. Floor cleaners, delivery robots, things like that, where they have a compressed cost of goods. So their whole cost of goods might be $500 or $600. So we built this to be able to provide to those people and to compete in some of the markets that are kind of really price suppressed. So, yeah, we’re excited about that. And the POE is I think very important for us. Having Power over Ethernet is a big breakthrough for us. So a lot of good stuff. But yeah, it’s just great walking around and seeing the applications.

 

Chris Matthieu: [00:03:41] I wanted to expand on that POE, the Power over Ethernet camera. It can go 100 meters powered by Ethernet with a camera or with a server in a closet somewhere. So in a production line, your conveyor belts or whatever, you could string these things out 100 meters far and do all your depth sensing way out.

 

Winn Hardin: [00:04:02] Yes. I mean, that’s super important. I mean, my mind’s been jumping from things to like, how do you deal with embedded systems versus traditional robotic applications or even machine vision? I mean, where do you see the most exciting part? But then you bring in the consumer high-volume applications that are basically transitioning machine vision capability down to the consumer electronics space. And I just kind of think you guys are busy as heck.

 

Chris Matthieu: [00:04:26] I mean, to your point, like when I think of our cameras, I’m thinking mostly robots. But if you look at that POE camera, it opens up tons of use cases for automation, for security and stores, for counting people and stuff in shopping malls.

 

Fred Angelopoulos: [00:04:43] Easy integration with existing industrial networks.

 

Chris Matthieu: [00:04:46] Yeah, and our SDK is open source and it runs on all of our product lines. So, I mean, you build something for one, it’ll work on any style of camera we’re shipping.

 

Fred Angelopoulos: [00:04:55] I think that’s the important part of our solution is the SDK. I think it’s the best out there, and it just makes it easy for the first-time integrator or manufacturer to to integrate RealSense.

 

Winn Hardin: [00:05:08] Yeah, absolutely. And it’s beautiful that you guys designed it with the intent to be compatible with ROS. Basic robotic operating system, open source.

 

Chris Matthieu: [00:05:15] We have a Python wrapper as well. So Python is like the new popular language for all the AI. You can just do Python to camera. And if you use Copilot or Cursor AI, it writes all the RealSense code for you. You don’t even have to be a programmer anymore to implement cool innovations with our products.

 

Fred Angelopoulos: [00:05:38] You still need those programmers to come up with the ideas.

 

Chris Matthieu: [00:05:40] Yes, but I think anyone now. So there’s like 30 million programmers in the world. Now I think with this type of Cursor/Copilot technology, anyone is a programmer now. Now it’s accessible. You have an idea, whether you can code or not, AI can write the code that does RealSense.

 

Winn Hardin: [00:05:56] Especially for point applications. You get into wide-area, large-scale enterprise stuff, it still requires a little bit of expertise.

 

Jimmy Carroll: [00:06:04] There’s a point that I keep bringing up a lot, and I do it a little bit too often, but I think it’s super relevant right now. But at the A3 Business Forum this past January, somebody was talking about automation, basically saying this technology’s been around for a long time, but in many ways we’re only just getting started. And one of the areas of growth that I’m looking at is small to medium-sized enterprise. And I’m just thinking about how an $80 camera that can offer 3D imaging capabilities along with easy-to-use cobots, flexible repurposeable cobots and cameras, can really start to open up new doors. Are you seeing anything there for some of these smaller companies?

 

Chris Matthieu: [00:06:46] Tons. Like Fred was saying, there’s so many specialty robots that are maybe inexpensive but very focused on a solution. I’ve seen our $80 cameras. I’ve seen our $250 cameras. It’s all based on the application or the price point they want. Our $80 camera it, one of the reasons it’s $80 is it doesn’t have an RGB on it. It’s just the depth sensors. I was like, I wonder if it can still do object detection without an RGB camera. So I asked Cursor AI: Would you help me out here? And sure enough it said, “Let’s use YOLO. I can detect a human. The human’s 3 meters away.” I’m like, “Holy cow!” Like, the only reason you need that RGB camera is if you’re trying to do something with color, like follow a red ball or something like that. Other than that, that $80 camera is going to suit you.

 

Winn Hardin: [00:07:35] I’ve seen those cameras. That capability has been around for a number of years. But we’re talking about $5,000 price points. And usually a little bit of a software commitment.

 

Chris Matthieu: [00:07:48] Yeah, another camera we don’t talk about at this show mostly, but the RealSense ID — that is taking off like wildfire: facial authentication. It’s an amazing new product that, again, gets us into a whole bunch of new markets: access control, security.

 

Winn Hardin: [00:08:06] You think about it. Ten, 15 years ago, biometrics freaked everybody out. Twenty years ago.

 

Chris Matthieu: [00:08:11] Now you got your phone. Everyone’s doing this all day long.

 

Fred Angelopoulos: [00:08:13] That was the thing, right? It’s the facial ID on your phone. Now it’s like, “Hey, I’m used to it. I need it. I can’t live without it.”

 

Winn Hardin: [00:08:22] I don’t want to type in that same thing 500 times a day.

 

Chris Matthieu: [00:08:25] Or I forgot my badge at work. Just use my face.

 

Winn Hardin: [00:08:29] Yeah, yeah ATMs. I mean, it’s the security actually.

 

Fred Angelopoulos: [00:08:33] Yeah, we’re working on a project right now in Korea and Vietnam with ATMs using facial authentication. And it’s really going to take off in the next few years because it’s like, I have a plastic badge. If I drop that badge in the parking lot, people are getting into the company. And we have so many applications. We have employees, some of our customers, employees that we have – 10% of them are blacklisted. But we can’t track them. So we can enter that data. And then the facial authentication will say, “Hey, you can’t enter.” But it’s really robotics everyone loves because it’s cool and you can see it and it’s doing amazing things. But everyone has a door.

 

Winn Hardin: [00:09:24] Does that help to maybe possibly blur the line between cobots and traditional industrial robots? Because now we have the ability to have even better human sensing, increased safety.

 

Fred Angelopoulos: [00:09:33] I talk with other robotic companies about that idea all the time. It’s like humanoids. I want my humanoid to only take commands or orders or missions from my family, and maybe I want it to bring me something instead of my wife something. So it has to know who people are around it, and our camera powers that.

 

Winn Hardin: [00:09:57] The humanoid applications, which is a growth area and a little bit of future state, but I could even see just that safety component. The biggest challenge, one of the biggest price points of a traditional robot workcell is the safety elements. And I know a lot of a number of different robotic arm manufacturers have started to use a multiple, multi-sensor approach to be able to provide, so you can go up higher on payloads and faster on speeds.

 

Chris Matthieu: [00:10:22] That’s an interesting topic. I think it was Figure, Figure Robots, they had a whole piece on that where with our cameras or in general you can do things 10x of a human. like perception, movement. So think about that in an automation environment, where a humanoid is 10x outperforming a human because of all the digital capabilities and perception.

 

Winn Hardin: [00:10:50] I would have liked some help last week when I was digging out some shrubs in the back. Yeah, 57 guys, I’m going to tell you. Swinging the maul’s not as much fun as it used to be.

 

Chris Matthieu: [00:10:59] One of our customers has as an application that does that actually using our technology. It’s driven by sunlight, solar, goes to identify: it’s a weed and not a weed. Picks the weed.

 

Winn Hardin: [00:11:13] Well, now we’re in agritech, which is one of the most exciting areas. Food production. It’s going to be such a critical component in the future. And automation has already been widely adopted in that space. But there’s a lot further to go, especially in terms of sustainability, environmental protection.

 

Chris Matthieu: [00:11:29] In that space we have customers that are using it for cattle to basically be able to understand their weight. And are they at the proper weight? Are they being well fed? Are they sick? I think there’s a big opportunity in agriculture.

 

Fred Angelopoulos: [00:11:49] I interviewed the founder of Greenzie at the Robotics Summit a couple of weeks ago. And they’re retrofitting commercial lawn mowers with RealSense cameras and auto navigation. So you don’t have to cut the grass anymore.

 

Winn Hardin: [00:12:05] Without running over the cat, which is probably an important design consideration.

 

Jimmy Carroll: [00:12:11] I’ve been following you guys for a long time. In a past life I wrote for a machine vision magazine. So it’s fun to be able to ask you guys questions. One of the big ones I have to ask, of course, is years ago, the news was put out there that Intel would stop production on RealSense, which was very surprising given how prolific – maybe not prolific but how widespread the product is used in robotics and elsewhere. And then the news came out that you’ll be spinning out. So talk a little bit about that. What led to this?

 

Chris Matthieu: [00:12:47] Yeah, sure. Well it’s interesting because going back to that, I think it was maybe 2018, 2019 when a statement went out that said Intel is winding down RealSense, and the fact was we were winding down the lidar line. So we’re just focusing on depth and not lidar. I think we have a better solution, better safety than lidar. But once once it’s on the internet, it’s got a life of its own. “I thought you guys shut down a line?” So we had to face that. It’s tough too because we have we have leading market share. I think we have probably 60% of all robotics. In the humanoid market, 80% at least. And the stuff that’s coming out of China is amazing. When you look at manufacturing, they need humanoids. Twenty-four-hour work, seven days a week, don’t call in sick. So Unitree, AgiBot, some of these companies are great partners of ours, and it’s amazing to see what they do. The Unitree stuff is spectacular.

 

Jimmy Carroll: [00:14:01] Yeah it’s cool.

 

Winn Hardin: [00:14:02] Yeah. OBX is doing a current trial with five different humanoid variants in their logistics warehouse, and we’ve had a quick conversation with them, but I want to dig a lot deeper with that. So you mentioned the mower, the auto-mower. And you mentioned seeing the cameras all across the show floor out there. Can you tell us a little bit more? Are they mainly for safety applications? This is for guidance optimization, robot placements? What other types?

 

Chris Matthieu: [00:14:31] Everything. When we walked through they were like on AMRs. They were using them like 360 degrees around an AMR, moving around the show. We saw them on forklifts. So as the forklift is coming, it can see, it could even volumize the pallet and have better positioning.

 

Winn Hardin: [00:14:51] A lot of path optimization and safety

 

Chris Matthieu: [00:14:56] Arms.

 

Fred Angelopoulos: [00:14:56] That’s a big part of it is the robotic arms. Because if you look at today’s world, you want to order something online. It’s late at night, and you want to get it the next morning. That’s because it’s robotic arms and AMRs doing that. So yeah, that’s a big part of it. Putting a camera in the hand of the arms so they can identify the right product to pick and place.

 

Fred Angelopoulos: [00:15:21] We have a little quadruped, a dog, in our booth. Walk around. It’s got a RealSense camera right above here. The Boston Dynamics robots, they’re walking the Intel fabs. I’ve seen them, and they’re listening to machines out of threshold, like frequencies. And they’re opening tickets, like support tickets, as they’re walking past the machine. It’s amazing.

 

Winn Hardin: [00:15:43] Especially with Boston Dynamics, for sure.

 

Chris Matthieu: [00:15:45] Yeah. And I gotta give some props to ANYbotics, in Switzerland and Zurich, because they have ANYmal X, which is another amazing robot. Everyone here loves to talk about Boston Dynamics.

 

Winn Hardin: [00:15:58] Because the YouTube videos are awesome.

 

Chris Matthieu: [00:16:03] But I mean ANYmal X goes into oil rigs in the middle of the sea, can climb steps.

 

Winn Hardin: [00:16:12] But can it do parkour?

 

Chris Matthieu: [00:16:14] They have wheels in their paws. And so as they’re going, if there’s time where they can just coast, the wheels recharge the batteries. So people are like, “Robot dogs are cool,” but they’re really important. They go into places that humans can’t go.

 

Winn Hardin: [00:16:31] Or shouldn’t.

 

Fred Angelopoulos: [00:16:33] While we’re giving props to our favorite robots, Agility Robotics with Digit is here at the show, and he’s got seven or more of our cameras in him. It’s the grasshopper leg-looking robot lifting all those boxes, the Amazons and stuff. That’s one of my favorites.

 

Chris Matthieu: [00:16:49] The cool thing is our customers come up with applications before we do in a lot of cases.

 

Winn Hardin: [00:16:56] Well, they’re the domain experts, right?

 

Chris Matthieu: [00:16:57] Absolutely, absolutely. There’s a company out of South San Francisco called Sembi. They were a startup. They use 16 of our cameras. They are going inside of supermarkets and doing real-time inventory while the customer is there. So they have to be aware of where the customer is, make sure it’s safe. But it’s an amazing application.

 

Fred Angelopoulos: [00:17:19] And kids think it’s cute. So they walk up to it and it’s got eyes and it knows how to be kind and gentle around them.

 

Winn Hardin: [00:17:27] That’s good, that’s good. So the next generation grows up. Well, we’re already seeing the concerns and the retrenchment against robotics and the general world – that’s changed.

 

Chris Matthieu: [00:17:35] Absolutely.

 

Winn Hardin: [00:17:35] Now AI is the new frontier of fear. But it’s all about education and familiarity.

 

Chris Matthieu: [00:17:42] Everyone wants a humanoid to do their dishes, fold their laundry, clean the house. They’re waiting.

 

Winn Hardin: [00:17:49] Watch my dog when I’m on the road. That would be super awesome. Boarding costs are killing me, killing me.

 

Fred Angelopoulos: [00:17:56] And there’s other applications that are incredibly unique that again we wouldn’t have thought of by using our technology. It’s not robotics space. We have one company that’s using our technology in grain silos. So instead of having to go in and measure, and grain silos can be a dangerous environment.

 

Winn Hardin: [00:18:13] A very dangerous environment.

 

Fred Angelopoulos: [00:18:14] So they just put the cameras in there and they can see the regression of the grain and can load this one and things like that. We have an application where, I don’t know if I can talk about the restaurant chains that are using it, but if you’re going to get a burrito or some Chinese food, and it’s one of these places where you slide down the line and they have the pans of food, and I’ll take this, take that. So if you run out of the chicken, what happens? The whole line stops. Now they have cameras on the pans. It’ll say like, “Oh, we’re below 50%,” and sends the message back to the kitchen: need to get chicken out here in the next five minutes. So those are the kind of things. Who’d have thought?

 

Winn Hardin: [00:18:59] I’m just happy to have buffets with warm food. That’s very cool.

 

Jimmy Carroll: [00:19:05] We sort of touched on this a little bit earlier, but the idea of there being a 3D camera with such a low price point. What’s been your experience with, I won’t say prototyping but like companies that are startups maybe being able to go from ideation to real-world deployment. You must work with a lot of companies in that.

 

Winn Hardin: [00:19:25] And what sort of guidance do you give them?

 

Chris Matthieu: [00:19:27] Yeah, those are great questions because this is really a big part of what RealSense is all about. We don’t just say, “Hey, here’s the cameras. Pick the one you want.” We take them through the whole process. So the first thing is like, oh, you guys got 30 SKUs. This is really about eight cameras. But whether you want them with the container or just on the board, but they don’t know where to start. What’s your application? What are you looking at? And we get them samples and we basically have weekly meetings with them as they develop. So we’re super supportive the whole way because we understand for a lot of these companies, they’re taking a leap of faith in us. And it’s like they’ve never done it before. And how do we optimize it? What’s the best way to do it?

 

Fred Angelopoulos: [00:20:10] No one’s ever done it before.

 

Chris Matthieu: [00:20:11] Exactly. So we pride ourselves on that. We collaborate very strongly with our customers and we have a really great relationship with them. We say it’s all about the customer.

 

Winn Hardin: [00:20:23] Sounds good. Warm and fuzzy.

 

Fred Angelopoulos: [00:20:25] Well, and Jimmy that’s another interesting point about makers and hackers building something in their garage with the $80 camera. You never know where that next big hit is going to come from. They’ll be calling us with $10,000, maybe higher-end cameras because it works on the $80 camera.

 

Jimmy Carroll: [00:20:41] Yeah. Yeah. Sure.

 

Winn Hardin: [00:20:45] So talk to us a little bit about how AI plays into the RealSense 3D product line. Is it important to you guys?

 

Chris Matthieu: [00:20:53] Extremely important. So a lot of our cameras have ASICs built in. So they’re doing a lot, all the depth on the camera. So I can run this on a Raspberry Pi with almost no compute, and it doesn’t even get hot because the RealSense camera is taking the brunt of all that, that AI and computation and calculations. We started some experiments that are starting to pan out in-house in our R&D group doing object detection with 3D instead of 2D. So like all the algorithms on the market, YOLOs, etc. They all do 2D, like webcam or RGB-type cam recognition. Well, if you have a doorbell, a smart doorbell, and you’re getting false alarms because there’s a raccoon in your yard instead of a human, if you use the 3D for AI, we’re finding that this is the next frontier of AI, I think, is moving from 2D to 3D AI.

 

Winn Hardin: [00:21:51] I wonder what that will mean for model efficiency. Model optimization, which which would be great. I think that’s one of the barriers to overcome for AI. So I need to design a good training model. I need the right data to be able to do so. But it’s hard to know whether you’re going down the right path. I mean, we see a lot of machine vision back in the day was all monochrome, and then color became another distinguishing feature that we could leverage. And then of course 3D in general. And now you guys are bringing that to the mass market.

 

Chris Matthieu: [00:22:20] I did an experiment and it worked. And it blew my mind. I said, can I just take a RealSense camera and stream all of its sensor data, telemetry, into ChatGPT or Claude or an LLM? And all these multimodal LLMs – again, no programming. You could just say, “Do you see a human in this stream?” And it’s like,”Yes.”

 

Fred Angelopoulos: [00:22:44] I was like, just ask it, “What do you see?”

 

Chris Matthieu: [00:22:46] Oh, I did that to my robot. I said, “What do you see?” And he looks at me and he goes, “I see an old man.” I’m not going to ask him that again, but then you just give it missions. You say, “If you don’t see a human, can you move around, look for a human, find a human, follow the human.” I did that at the Robotics Summit in 10 minutes. I talked to an LLM, and it wrote all the code for my RealSense ROS robot to follow me in 10 minutes.

 

Winn Hardin: [00:23:15] Wow.

 

Chris Matthieu: [00:23:15] I didn’t write a line, but it was the ChatGPT or the LLM saying, “I see a human,” translating that into distance, and saying ROS twist commands to move the robot, stop a meter away. The crowd was just like, “Oh my gosh.” Like this opens up even more for everybody. Like, this is robotics for everybody.

 

Winn Hardin: [00:23:33] Yeah, a lot of people out there are wondering what is the real impact of AI. And you guys are talking about it directly right now. I mean, we do a lot of programming in our business and everything. So we’ve leveraged that for plug-in, widget optimization, whatever it is. But, I mean, end-to-end robotic applications.

 

Chris Matthieu: [00:23:50] If you think about it, it’s like a brain. You just gave that robot a brain. And you’ve given it a mission. So it’s figuring everything out and sending ROS commands to move the robot. If you think about what’s happening there, it’s mind-blowing. And like you said earlier, this is the beginning. Everything is iterating.

 

Winn Hardin: [00:24:08] Beginning of the beginning.

 

Jimmy Carroll: [00:24:09] The intersection of AI and 3D is obviously very interesting, but I imagine you see some use in food sorting and food production in general. Because you’re having to deal with organic, everything’s slightly different, a chicken breast, for example, or oranges or something like that.

 

Fred Angelopoulos: [00:24:35] Yeah. We have an application with strawberry farms where the device has to be able to tell which is ripe, which is underripe, which is overripe. So you have a 33% chance of getting the right one. But the AI vision, it’s 99.99% you’re getting the right one.

 

Chris Matthieu: [00:24:54] There was a student at the Robotics Summit that in two weeks he built a tomato picker using RealSense. And the thing just had wheels and moved down a track, and it would look at every tomato. Kind of like you’re saying. It looked for: Is it bruised? Is it ripe? And it actually had the little thing that would twist the tomato off and put it in a sack. I did that in two weeks: the end-to-end robot.

 

Winn Hardin: [00:25:16] Similar applications on apple picking and tomatoes 10 to 15 years ago. Back in the day when we were working and publishing so much, and there was a Dutch company that was doing at the time from a government grant, and they had spent years trying to solve the agriculture automated picker application. And in the end, it still wasn’t very good because the systems weren’t ready for ambient light, changes, making it so difficult.

 

Chris Matthieu: [00:25:39] The 3D element takes all that, timing. It’s all happening right now. And everyone’s open sourcing things, iterating on top of each other. And so it’s getting easier faster. I like easy.

 

Jimmy Carroll: [00:25:52] And RealSense to me seems like something that could also be used in different aspects of different processes. And what I mean specifically is like we’re talking about AI, right? And digital twins is something that comes up a lot with AI lately. And given that the RealSense is so inexpensive, are you finding a lot of people are using RealSense to create digital twins, do scans?

 

Fred Angelopoulos: [00:26:14] Yeah, absolutely. It’s probably one of our first customers in that area was Rihanna. She was doing her own clothing line, and she wanted people to be able to create a digital twin and try her stuff on. So yeah, that was really one of our first applications about three years ago on digital twins. And now we see it a lot. Fit, Match, Shape, Scale. These are companies that are using scanning technology and digital twins to compare your health, basically your fitness, areas that you need to work on, areas that are in good shape and things like that. So, yeah, digital twins is a big part of what we do. And I think, again, it makes the experience much more satisfying for the user.

 

Chris Matthieu: [00:27:00] We released an Android edition of our SDK. So now people are putting the scanners on their Android tablets and creating the digital twins walking around with the scanners.

 

Winn Hardin: [00:27:11] If you’re a manufacturer or an industrial designer, say you have a digital twin. So we’ve got a simulated environment of a production line of the plant. To be able to connect that to real-time operations and be able to visualize the whole thing, we’ve either got to install miles and miles of cable and maybe hundreds or more individual point sensors. We were just talking with another very interesting group, Spot AI, earlier and talking about: Can one camera replace five, 10, 20 point sensors? If we’re just looking at conveyor speeds. If we’re looking at hands-off. We don’t need to have the photo eye and the accumulator. We can just visually detect all of that type of information. To me that seems like it can be a real game-changer.

 

Chris Matthieu: [00:27:59] Totally. You’ve got the RGB. It’s a great RGB camera in the device. So you could just look at things and do your object detection and use depth, whatever distance you need to set it at, identify where it’s at.

 

Winn Hardin: [00:28:14] What’s next for you guys? What are you excited about?

 

Fred Angelopoulos: [00:28:32] I think just excited for the future. We are spinning out of Intel over the next few months, and I think Intel’s been an incredible partner, and they’re going to be a key investor for us. But we’re excited to go on our own path here, giving us maybe the freedom to tell our story and kind of run on our own. So we’re excited about that. We’ve got some new technologies coming in, coming on board in the next 12 to 18 months around safety and things like that that I think are going to make a huge impact on the industry. So I think that’s an area that still needs to be addressed because it’s an environment where you have robots and humans.

 

Winn Hardin: [00:29:18] Is this mainly a software envelope development, application-specific type of stuff?

 

Chris Matthieu: [00:29:22] It’s a combination of software and hardware. But a big part of that is going to be software.

 

Winn Hardin: [00:29:26] Absolutely. Absolutely. Chris, Fred, it’s really been a pleasure. Thank you so much for joining us today. And I’m sure we’re going to see you a lot more on the road. I think our paths will be intersecting. As always, thanks to the guys. Thanks to Jimmy. Thanks for joining us. If you’ve got any questions for RealSense, please put them below. We’ll make sure it gets to the guys. If you want to see any past episodes, go to manufacturing-matters.com, where as always we’re talking about the technology and trends that are reshaping the global manufacturing industry. You can also find us on all your favorite podcast platforms. So until next time, thanks again guys.

 

Chris Matthieu: [00:29:59] Thank you so much.

 

Fred Angelopoulos: [00:30:00] Thanks everybody.