[MUSIC] The scanner in this video is the NextEngine 3D scanner which is another structured light scanner but it uses a different technology than the Kinects. It'll still be projecting a pattern onto the object and interpreting the shape of the objects based on its distortion. But in this case, it's using a red laser line sweeping across one side of the object. At $3,000 it's a little out of reach of the hobbyist, but a lot of universities and a lot of fab labs are going to have one, and it's a really good example of what's possible at the upper end of the spectrum. This is a much higher resolution scanner than the Kinect, a bit more of entertainment-style 3D scanning. It uses structured light scanning, similarly to the Kinect. But instead of getting an instantaneous scan of the whole object, this slowly takes a scan of one side at a time. So this is kind of, the opposite of real time 3D scanning. For one side of the object, it could take anywhere between two and four minutes, depending on the detail that you want out of it. And to get a full 360 degrees of the object, you then reposition that object to scan the next side. It might take a couple hours of your time to get a good 3D scan but the result is going to be higher resolution and it's also going to be dimensionally accurate. Any of these processes you have to think about, how reflective the material is to the color of light that matters. So the red laser light sweeping across the object. The computer needs to be able to see that solid line move over the shape of the object. If you have something like a silver pocket watch, I was really excited to get a 3D scan of this but there's no way it's going to work because it's shiny. And the laser hits this silver and just kind of bends and reflects in all different ways. So if the laser can't register a solid red line, it's not going to take any measurements. For the demonstration, I'm using my antique adding machine. It's a mechanical calculator. [SOUND] And I'm really happy it's still in working condition, and I want to make a high-resolution 3D model of it. It's a good material for it, it's kind of a matte, dull metal. The software that we'll use comes with the scanner. You don't need anything extra. And it's pretty user friendly. You kind of hit one button after another, you don't have to solve a lot of problems with this one. It's just important to follow its recommendations, and measure things accurately, and not do a lot of guesswork. So to demonstrate how that works, I'm going to hit Start. So it's going to start a new scan. The scanner's going to light up. The first step is to get the object in the view of the scanner. So it gives you a video preview of what the camera can see, so you can set your object in the middle of the frame. That gives you the option of doing a quick scan of low resolution or a high definition scan with higher resolution. So you can see that it's going to measure anywhere between 55 points for cubic inch all the way up to 4,500 points for cubic inch. The main difference is like that can take just a minute and a half. Highest resolution will take three and a half minutes to scan the entire surface. But I want a high resolution scan so I'm going to give it a shot. The other option that you have to be concerned with is the range. You can either choose between a macro, or a wide range, or an extended, a little bit further away range. But you choose one of these three and put your object that prescribes distance. So for this object, it's kind of a medium sized object. So I'll use the wide range and it tells me that the ideal distance from the sensor is 17 inches and a maximum of 22 inches. So I have my yard stick and I'm going to measure that out so I don't want to guess about that. So ideal 17 inches, it can be a couple of inches closer than that, a couple of inches further away, 22 there, so this is a good distance. Measure it and get it in the middle of the frame. So at this point, I want to start here, I want to start a sideview, but then I'll center it and hit Go. And the rest of it is pretty automatic, it will take a picture of the objects. And try out a few different lenses and so it finds the right focal link. And then, it will start the laser scanning process. It's grabbing thousands and thousands of data points. So after the scan, it's going to take several minutes of processing. Even on a modern computer like a laptop or an i7 and a video game graphics card, it's still going to take several minutes, maybe even 10 to 20 minutes on the high resolution scans, to process just that one side. The real advantage of this is that it doesn't make up data, it doesn't guess. So with the photographic data or the connect data, it's doing it's best to create a mesh from the object in front of it. But those technologies will have a tendency to fill in gaps and sort of blur things together if it can't quite tell what's going on. With the laser scanner, if the laser reflected in a way that is going to give it junk data, it'll ignore that data. It's not going to try to give me a match that's not going to do it's best guess. Is that they going to give me a surface or it's going to ignore the surface? So that's really important for engineering applications if it didn't get an accurate result, then you don't want it to give you a guess. So you can see right away, it grabbed a lot of detail but it also missed a lot of information, too. This is why this is such a so processing, there are a little bit of detail per scan. But we have to turn the object a little bit, scan again and it will be a process of aligning those scans after the fact. I've scan this before and the result is a really detailed down so like the pattern on here which is kind of an embossed or a stamped metal, the 3D scanner is able to capture that. So this is a really great technology for getting dimensionally accurate scans, really high detail. It can be a very time consuming process but that is the cost of getting a really high resolution scan. [MUSIC] [SOUND]
