Laser-Scanning Workflows for Construction Sites: Reality Capture, Modeling, and BIM

JACOB RASKIN: Hey, good morning. Can everybody hear me? Terrific. I'm Jacob Raskin, Skanska USA in New York City. I am full time on a job site. And I help them run their laser scanning. And hopefully, that's why you guys are here this morning. Hopefully, everybody's had a little bit of coffee, shaking off last night's activities. Pretty good, almost there. I'm going to let some of these new folks filter in real quick.

But while they're filtering in, don't sit down too quick. We at Skanska have this lovely tradition. Every day, we start the morning with a little thing we call stretch and flex. So while you still have some room between you and your neighbors next to you, why doesn't everybody get up real quick? We're going to shake out the bones, wake up the mind a little bit.

So let's start with a little Jazzercise and do a little back and forth real quick. Get the legs awake in case the coffee isn't there yet. Let's give it five more seconds. Looking good. OK, great. Let's go up towards the sky. Those of us wearing belts, almost there. It works easier in jeans in the field. OK, terrific. I'll give us the short cut one. Let's do the neck roll, going slowly around. Don't speed it. Be gentle on yourself. It is only 8:00 AM after all. And let's go gently the other way.

All right. You guys are looking good. A couple more. Let's do this forearm reach. You can do two at once. You can stretch out that back leg and reach forward with this. You pull the fingers back here. And now we're going to flip them over, pull them down. And then you turn it out. Go the other way. All right, we're going to switch sides. Everybody's looking great. I love the participation. Pulling back and up, left leg is reaching back. Great. Going down, pulling the fingers back, getting the top of the forearm, and now we're going back out from behind you.

All right, give yourselves a nice big round of applause. Excellent participation. Looking great, feeling even better, love the enthusiasm. Thanks so much for everyone being with me this morning. Really glad to see such a large crowd coming out here. It really means a lot to get this sort of participation and energy coming in here.

So let's go over the agenda. You've already done step one. Congratulations. 8:01, we're already flying. We're going to talk about the laser scanner itself. We're going to talk about some best practices of how to run it effectively on the job site. It is just a tool after all, so a lot of using it correctly is knowing how to get it set up correctly and operating correctly on the job site. And that's a lot easier said than done in my experience. And I think that it might be helpful for a lot of us to review that.

We're going to go over processing techniques. What happens after you come back from the field, and you've got this big heap of data, then what? We're going to learn to explore the scan or refresh it, if you guys are already familiar. And then we're going to get into what I love the most, where you can pull the scan into some design softwares. I'll show examples with AutoCAD and Revit softwares today.

And then we're going to jump into another software suite called ReCap, which is really helpful for viewing the laser scan product, the point cloud. And after that, I'd be happy to take some questions. I've got a fair amount of material to cover today, so I'd love it if we can save questions for the end. But I like the energy so far.

All right, let's talk about the basics. This is a laser scanner. The unit itself is what's sitting on top of the tripod. The tripod, though, will frequently come with it. But it bears reminding that they're not the same thing, and most laser scanners are kind of the bread box that sits on top. And they're separate objects.

What it does is it's like a digital tape measure. A tape measure traditionally would look across the room and say, OK, from A to B in one dimension, this is how far things are. And this has taken this step further, and it will look in all directions. And after you let it run, it will give you this collection of measurements, rather than just one measurement in one direction. And that's the main advantage of jumping to this from a point and shoot laser tape measure or a tape measure itself.

The product-- what it spits out to you, what you're hoping to gain from it in the end-- it's called a point cloud. This is a view of a point cloud right here, sort of inside the space, looking out away from it. And it's a collection of x, y, and z-coordinates. And each one of those points also has some other properties attached to it. So what we can see here, for example, is that it has some black and white values assigned to it as well. And that gives it some texture beyond just the geometry values that they carry.

So let's talk about some basics and good stuff to remember when you run your laser scanning on your construction job site, where I play in the sand most days. The first thing you want to think about is, what's your actual objective? So you say you want to go out and laser scan the site or the existing conditions. But you've got to think about first, what is it I really want to learn about?

Some of these questions on the board right here might help prompt some thinking about, OK, I want to go out, and I might want to measure all the steel in the space. Maybe it's a renovation job, and there's a tremendous number of existing conditions that I need to know about before I get my construction folks in there to get started. So that I can tell them, OK, you're going to encounter some low points here, and then you're going to encounter some high points back there in the structural stuff. So look out for that. And you know to dip under that.

Things like this will help you guide where you're going to send the laser scanner as you're walking around the site with it, doing your scanning. Or maybe you don't need to go around the entire site at all. Maybe you are only interested, for example, in just looking at one particular area that's a really key feature.

Another thing to think about is how much time can you allot to the laser scanning. To do it correctly and get the most value out of it, you do need to give it some time to run. This can sometimes take multiple days. You might come back for an entire week, day after day. It's very infrequent that you can really get all of what you really want to know by just parking it once, letting it spin around, and saying OK, we're done. We did it. And you won't really know how well you did it, unfortunately, until you've gone back to the computer, and done the processing technique, and opened it up, and kind of said, OK, let's evaluate how complete of a picture did we really gather about the site conditions when we went out that day.

It brings us to how much time is enough then. So, as with many things, there's a trade-off between how much time you're going to spend doing it and the general quality of the product. If you go fast, you can really only ever expect to get a little pretty high quality data or a lot of relatively mediocre data. Because you sped through and went very quickly from point to point as you went across your site to scan.

So this is one of the planning things that you need to keep in mind before you ever even get the thing out of the box, and put it on the tripod, and send people out to the field. It's just to say, OK, I expect that we can spend eight hours times 5 days taking in site conditions, or it's a very protected site. There's rules about going in, and they're only going to let us in for one day. So it's a one hit wonder. So ask as many questions as you can before you bring the toy in and set it up to run. Because you're only going to have that limited window to take in exactly what you think the most important thing is. And the rest of it, you'll either have to reference existing documents or ask for another chance some point later.

Let's talk about some basics of what else comes with a laser scanner. What else should you be doing on site, aside from just parking it and hitting run? You want to set up some targets for it to try to see. And the reason you want to do this is because when you run the post-processing, the algorithms that come in the software package with your laser scanning kit will be trained to look for these sorts of targets. And there's typically two types. One is a sphere, which is a pretty perfectly round white ball, and the other is a checkerboard. And what these software algorithms are trained to do for you, which is great, is to look for round things that look like spheres and trained to look for the crosshairs of the checkerboard.

And so you want to set up your site before you hit Run on your scanner with as best ability to look for these things as you can get. You want to affix these targets in a very permanent fashion to the wall, or at least as permanent as your site will allow you. Sometimes you're going to go into, say, an operating hospital. It's not very likely that they're going to let you tape up a checkerboard target to the wall. There are some workarounds for this. Some of them can be printed on magnetic strips. And they can be taped up for the time that you're going to do your scanning and then removed later.

But for the most part, if you're going to do multi-day scanning, you really want to try to take every effort and get your client to buy into letting you have those targets up for all of the time that you expect to be doing your scan project. The spheres are a little more negotiable. They go up easily. They come down easily. But the general theory is you never want to move them. If you're going to do a bunch of scans, the targets have to stay exactly where they started.

Once you've gone out of a space-- you go around the corner. Two hallways intersect, and you come up this one, go across. Once it's out of sight, then that's fine. It doesn't matter as much. And you can borrow the spheres and move them around with you. But the checkerboards, they go up, they stay up. So just run off a lot, bring a lot of tape, or glue, or whatever you can tape it to a wall or column, and expect them to be there for as long as possible. You're going to get extremely unlucky, unfun results if they start moving around.

And there's horror stories out there in the wild about people who have thought they were doing the right thing by picking up a fallen tag off the ground. And they put it back up on the wall, and they say, oh, OK, it looks pretty close. I can even see the dust mark where I can-- I know exactly where it was. But these scanners are so precise that they really know down to-- way better than 1/16 of an inch about where these checkers are. And you'll have these scans not aligning properly. But the program thinks it's doing the right thing. It's just not going to be a very good time for your operator, who's going to be putting these together in the end.

While you're going around site, you're going to encounter some obstructions, unless you're working in a really lovely wide open warehouse. But that's infrequent. Most of the time, we're scanning to gain the really interesting features somewhere on site. You're going to be passing by doors. You're going to be walking down hallways. You might even be looking through windows. You might be looking into a mirror. And all of these features throw off the scanner in a slightly different way. A window has a little bit of refraction as light passes through it. So it will look through it. But the readings that you might get from whatever is on the other side of the window are not going to be accurate because it's going to wave a little bit as it goes through the window.

Mirrors are tricky. It will sort of read into the space and read the visual projection of what it saw in the mirror. So it's almost as if you'll be seeing into another room that doesn't exist at all on the other side of your wall. Puddles can be challenging. It'll look into a puddle and sort of see things. Water is similar to glass. It has a diffraction property different than air. So your picture comes out a little wonky.

As you're going across your site, you want to say, OK, I want to maximize my efficiency. I want to make the most use of my time. And you just need to be aware of what your particular unit is capable of reaching. How far away is it going to look and still be within a good accuracy? And you want some overlap, sort of like what I'm displaying here with the green dots. This was a planning chart that I set up for one of our journeys out into the site to try to tell the operator, this is roughly where I want you to stand the scanner to get enough overlap between where I know the targets will be.

And so that's another step that you need to take before you ever turn the thing on and start running, is to say, OK, where? I know how much time I have now. So how far in between do I want to go? Again, thinking, the faster that you go and the further spread out, the less likely that you are to have an adequate amount of stuff in the middle bleeding between one scan and the next to get them to talk nicely together and the more challenges you may experience further on when you do the processing.

Which brings us right to the processing. So you've gone around in your site. Someone's parked an SD card on your desk and said, here you go, boss. And they walk away. Now what? So now you get to run it through the software suite that comes with your scanner. It might be [? Laika. ?] It might be FARO. The images that I'm showing you here happen to come from FARO. This program's called FARO SCENE. That's just what we use on my job site. So those are the images I have for you today. No preference either way beyond that.

And so when you open it up, it's first going to try to chew through all of this. It's going to do some rehashing, let's say, of this database. That is what the scanner saw on that day. And it's going to turn it into a whole bunch of representations of those scans, 3D space, and each one of these colors represents a different scan that it saw. And it's trying to mesh them all together in a nice way. But it's not always easy to basically go from the top image to the bottom image, where everything suddenly says, oh, OK, yeah, this makes a lot of sense. I can see the steel, and the steel lines up with the steel. And the floor lines up with the floor, and the walls are all straight. It makes a lot of sense. Not necessarily exactly what happens the first time you run it through the automation cycle.

So there are some techniques that we use. And the first one we're going to talk about is called cloud to cloud. And this is a pretty nice technology that exists, where it's basically all in the computer. And it's basically just looking to say, do I see some distinct feature in one scan? And can I see that same looking distinct feature in another scan? And if so, I'll line them up, and I'll assume that these two parts match. And now I have a relatable distance between these two scans and an angle about which to lock them together and say, OK, I know that A lines up with B. And it's pretty good. And you can go through levels of refinement to get better and better closeness out of your scans.

But it's not always the most accurate thing. It's also very time intensive to let it run on its own without any help. There is another technique called target based, which is why you hung all those targets out in the field to begin with. And again, that automation software is going to look for all the targets that you should have placed nicely through the site. And it's going to say, OK, great. I don't need to see a pattern in the points themselves in the geometry. All I need now is to look for a pattern in space between targets A, B, C, and D. And now I know that all those points match up together. And that's a lot faster because it's a lot fewer things to look through.

However, it's hugely contingent on your targets being nicely distributed for the first time. So again, going back to the field and the preplanning, making sure that all your targets are being hung nicely in the places that you think they'll be seen, so that when you get to this point, you can just go in here and say, open it up. OK, yes, I see that target that I hung on the wall. I see the target that I hung on the column, and I see the sphere that I placed on the ground in between. It goes a lot faster, but it just requires more preplanning and a little bit of verification to make sure that, yes, everything that I thought I could see on that day is actually still there when I open it up.

And it's not perfect, unfortunately. Sometimes it gets a little overzealous. This is a picture from the train station, where I'm working right now. And we went down onto the tracks a couple of weeks ago, and someone was asking about the steel. So we went down to look at the steel. And they said, OK, yeah, we even took the targets for you down. So here you go. This should be nice and quick. And I ran it through. And it blew up and it said, I quit. This is partially why.

It accidentally recognized all of the rivets as spheres because it's trained to look for a round feature of a certain diameter. By shear dumb luck, all of these steel elements fit those properties. So it drew a whole bunch of false positives for those spheres. It drew some other false positives for checkerboard targets. That's these little crosshairs right here. There's no way that the people got 40 feet up in the air to tape the checkerboard to the underside of that girder, though. So you can go through and help the program out and delete these false positives.

It gets a little better, and sometimes you get some false negatives, too. The checkerboards work best when you're face on with them, when you're looking pretty much directly at it. And what I also learned the hard way is that it also helps when your printer still has a full toner in it. So my guy did exactly what I wanted. He taped up all my targets in beautiful locations. The scanner could see all of them. But the funny thing was that the algorithm couldn't figure it out because it said no, I see a big black box here and only a small one here. Because the toner quit on the first half of each sheet. So, best laid plans.

Another way to do it is if your unit is enabled with total station capabilities or can otherwise link to a GPS signal, you can use that to bake that into your scan. And your scan will basically be right where it should be, at least to some GPS system, automatically when you open it up. Those units tend to be more expensive, though. So your mileage may vary as to what do you think you're getting-- just another consideration when you're commissioning or looking to purchase a unit.

Roughly an hour of time in the field is worth an hour of time in the office. Speaking with some other people who do this frequently for construction projects or planning services, I've found that that's a number that holds true pretty well. It's not hard and fast. There are different things that will vary again. You expected targets to run quickly, but you've got a bunch of false positives or negatives, so you had to go in and do some surgery on it. OK, so you had to spend some more time. But you didn't know that until the toner quit on the printer, and that was your experience. So it's just a good planning number to keep that in mind.

But more than anything, the more preplanning you can do and plan for in the field and set everything up nicely before the scanner runs, the better your experience is likely to be when you get to the post-processing time back in the office. OK. We're getting closer. You've baked the scan. The thing says, OK, I registered beautifully. And now we're ready to run. So in the Autodesk world, your program is called a ReCap. And this program is what you can use to step inside those laser scans and go back to the site that day.

And it has some great capabilities. You can zoom, and pan, and orbit, and fly through the point cloud. And you can also measure between all the points with a couple of different locks. It's not just point to point. You can clean the scan. It lets you delete some of the points if you've got, say, like that mirror example we spoke about earlier. You say, OK, I know there's nothing on the other side of that wall that I care about. So I'm going to take a little marquee select, and sweep those points away, and hit Delete. And now they cease to appear.

And it's often helpful when you're transmitting a laser scan file to some partners to clean up the scan. Because there's a lot of extra data in there-- just points that you're not particularly interested in, or maybe whoever you're dealing with who needs the scan today, they're only interested in, say, the steel. So you could even just cut out the entire bottom part if all they want to do is investigate the top of what you saw that day. And the next day, you might cut out just the top and say here's a floor flatness study for someone else. And so that's one of the things you can do with this program to help speed people up as they're investigating these as built conditions.

There's a couple of different ways to view the scan. We were just viewing what's called the RGB value. I happened to run that scan in black and white mode because it goes faster. So what you're really seeing here is basically an intensity reading, although it tells you, this is my black and white value. They're actually one and the same when you run in black and white. But you can view it in a couple of different ways. And it's not contingent on what color or black and white settings you used in the field that day. This is an intensity read, which gives you kind of that color bar associated color to how well did it read any given point, or any given surface, really.

So there were some general condition that it read down on the floor. The walls had a different pattern, and these shiny conduits up on the top-- different again. Sometimes it's just helpful if you're just presented a space and said, OK, look, you can detect different surfaces by using these different colors to help see what's going on or see perhaps what the scanner missed. Just because you stood out there and looked at it doesn't mean it actually grabbed everything. Sometimes there are dark spots. Sometimes a bird flies in front of just what you wanted to see.

Another way to view it-- sorry, wrong way here-- is the elevation view. This can be really helpful not necessarily to say, OK, obviously, here in the whole space, the floor is low, and the ceiling is high. What we can do is you can refine the limits of that gradient, and you can get down to a very limited tolerance and say, I want to study the floor. And I want to know how close to the average or to the theoretical value of that floor elevation did the scanner actually read. And you can see the high points and the low points.

And maybe if something is so far out of tolerance, it actually ceases to be on your gradient at all, because it's exceeded the bounds. And so there's some good stories out there of people using this against a curtain wall system to try to look at the tolerance of how well the curtains were actually mounted one panel to a panel to panel. Turns out they may not be actually all that flush. But you can go back to that contractor now and say, hey, I know the window is physically on the building. But you're so far out of tolerance, you're actually at risk of letting water flow into your building.

There's an entire other way to view the scan, though. It doesn't just have to be in the points. And what you can do is you can view the photo sphere. And frequently, what these scanners are doing almost entirely for the models that you're probably encountering in this day is it's taking a 360 photo effectively while it's reading in the points. And that's how it's assigning those RGB values, or those black and white intensity values, to the points themselves. And so right here, I kind of tried to set up a slider view between the right half of the screen and the left half of the screen, switching one between the other.

And what you get when you step into this photo sphere view is a very kind of Google Street View experience of what you saw on your site that day. And the photo quality is really pretty high. But the really good stuff that I like about this experience is that you can measure starting in the photo, switch back to the points to make sure that you're on exactly the right point that you thought you were on, and then go back to the photo and take a screenshot. And I can say I know I'm very confident when I tell you that from that floor to the bottom of this girder hiding behind this column here, it's exactly 11' 3.5".

And you can say, OK, great. I didn't just have to make that [? blooby ?] mark up on top of some site photo. I can spin this around and say with pretty 99% certainty, I know exactly how high that piece of steel is above the floor. There's some limitations in the photos for your view, though. It's not perfect. It doesn't give you as many capabilities as when you're walking around in the points. You can only step into a photo sphere where the scanner actually stood that day.

Google Street View is amazing because they take photos every 5 or so feet down the road. Very infrequently do we have time to do that in the construction enterprise, but it would be lovely if we could scan a site like that and say, I can tell you exactly how good it is every 5 feet. I've never had that kind of time, though. And you may not have the experience either. So you're limited to basically zooming and panning, but you can start that measurement or finish that measurement within the photo sphere. And you can pan around. You'll get this nice floor map in the photo sphere view. And that blue cone is your view angle. And it's telling you this is what you are looking at right now. So it gives you some nice indicators and some nice cues to say, OK, what am I looking at? Where am I?

Here's an example of taking a measurement. We're measuring from the steel, where that point is right here. And we're measuring down to the floor. And this little white dot right here is telling you basically the orthogonal plane to whatever point it's landing on. And it's a nice cue to say, OK, I think I'm on the floor. Does the scan also think I'm on the floor? And sometimes yes, and sometimes no. And you say, OK, this is a pretty flat floor. But the scan may not think that. It may think, oh, I see all these ridges. And you'll see the white circle kind of dancing along as you track your mouse around, trying to find just the right place to land that measurement.

There's some other nice measurement features. In case you haven't looked in ReCap recently, I think they added this in the version 4. You can now measure rounds. And it will read you some properties about the round object that you've landed on. So right here, what this measurement is measuring is between the soda bottle and the air can, between the outer surfaces of those two. But you can get some other statistics about some round objects. And you don't need to measure between them just to get them. You just say, I want to look at this pipe and measure the pipe for itself.

And it will tell you the diameter, and it's also telling me basically the angular elevation of which way is this circle pointing, so to speak. Where's the axis of this pipe going? And it's nice because you can snap two points on a pipe, for example, and say, OK, I will measure here, and I'll measure here. And I look at the delta z between them. And I can suddenly determine the pitch. And I never had to be on site with the level to say, OK, do I think this is at an eighth, or a quarter, or zero? Now you have proof.

And again, here's another example of measuring in just one dimension. I happened to pick z, but you can also lock it in x and y. You just need to be cognizant that if your scan is not referenced to any system, you may not have anything really of value to say, I want to know only in x or only in y. Those axes may not be aligned to anything. So it's just good things to remember how much information did I get fed in before I started reading answers out.

Let's look at how we can now take this scan, now that we've baked it, and get it into some design software. I'll show you two examples today of going into AutoCAD and going into Revit. So what I like to do in CAD is start with the background. Let's say I work with the general contractor. So I borrow one from my designer from my architect. And that CAD file has a zero point somewhere in it. And I get fed some survey coordinate information from my survey team that works with me on site. And I'm expecting that point cloud to jump in right where I think it should.

So there's a couple of fine tunings that you just need to do to make sure that that happens. It treats it just like an x reference. For those of you familiar with AutoCAD, you do Insert. And then there is an Attach Point Cloud. They give you its own dedicated quick button. And then you just need to be sure that where you are telling CAD to insert it is where you think it should be. And this is the key. And sometimes a checked box is still checked from the last thing that you did. And that's just the one thing. And there's horror stories of things flying apart just because of a little checkbox said, oh, well, put it where the last guy had the settings to pull into CAD.

But assuming that you're being fed good survey coordinates and the survey coordinates from your field verification team are matching up to, let's say, what your designer's AutoCAD coordinates are, you should be able to set everything to 0 and just say, put it in at zero. The zero points are going to be the same. The elevation should be roughly the same. And you can tell it, I don't need to do any input to this. And you should basically be able to go from blank canvas to full colors on the canvas with two clicks. I can't guarantee you that. Obviously, you need to know your site and your conditions better than me. But this is possible to basically be able to go right into it and say Insert, and there it is.

You can get a similar workflow going for Revit, not much different. They have this point cloud attach method, and it treats it roughly the same. The one trick in Revit, though, is just to know where the survey point and base point are. And those, as the Revit users in the room will remember, are generally hidden from the user because its best practice is not to ever touch them. If, however, your Revit project does have a survey point different from the base point, it's just good to remind yourself where are those.

Because it might come to light that if you do by shared coordinates, that might be the key thing to get your point to land right where you think it should be. But if you don't know that about your project yet, you might have to do a little experimentation to find out exactly where the base point is or where the base point in your cloud was. But you can expect similar results, and it should land right where you think it should.

The next key feature that I've really derived benefit from in my experience on site is being able to overlay these point clouds in a Navisworks environment and bringing Navis files into ReCap. That's another recent feature in case you haven't spun around in ReCap recently. So what you're seeing here are two examples of bringing Navis elements-- sorry, no, that's the Navis up there, and this is Navis down here. So this is zoomed in quite close to an MEP model overlaid on the points that they were modeled off of. And this is some new construction being modeled on top of my architect's model.

So let's start with the ReCap example. You go into Real View. And there's a button called Attach Project. First, you need to set up an NWD file from your Navis environment, and you need to have a lot of show and hide pretty well filtered out. You really only want to be showing in that NWD only the things you really want to see overlaid against your photo sphere. You can only view this in the photo sphere view. Again, that's why you have to start in the real view. It doesn't work over the points unfortunately in ReCap. Sorry about that.

But once you get all your show and hide set up, you can really generate these lovely images that show you exactly where all of your future construction, for example, is expected to be landing on top of your existing conditions. So on my site, for example, we have some really deep steel elements. That's what you're seeing here. And I'll go back and I'll show the first image. That's the girder right there before we did any work on it. And this is what we're expecting to do to it. And we're going to cut a nice hole in it and run our mechanical rack and a whole bunch of other flavors of MEP through it.

And this is a really great talking point to be able to hand to people, even though it's not dimensioned to anything, and I'm stuck standing here, looking at what I'm looking at. But it's really helpful to be able to say to maybe somebody on site and say, what is it I'm actually doing. Why am I cutting this hole? Well, this is why, sir. It's because we've got to run our mechanical pipe rack through it.

Helpful things in my experience just to remember when you're going to try to do a representation like this-- if you're being given or if you are the operator in charge of a Revit file with a whole bunch of architectural features in it, you just want to hide things like the ceiling. Even the see-through ceiling to hold a light on it will show up here as kind of a gray plane. And so just remembering to filter out a lot of elements-- like here, for example, I've filtered out walls, definitely ceilings. Usually light fixtures usually good stuff to take away and just hide in that NWD Save before you try to represent it here. So here's the comparison one more time.

The other way that you can view things is going back into Navisworks and pulling in a ReCap file. Navisworks recently has been able to open up a tremendous number of file formats, and ReCap is one of the formats that it now supports. So it's lovely. You can't represent the photo sphere itself in Navis. You can't spin around in it. That's for ReCap. But you can represent all the points, and all those points will come in with that RBG or intensity value attached to it. And so let's tighten here, and I want to study this tight spot where I've got some existing facilities. Those are the black elements, and I've got some new stuff that I want to put up nearby. Those are the colorful MEP elements here.

So what you can do now is you can now verify and say, OK, somebody handed me the previous phase's construction model, which I painted black. Then I went out and laser scanned it to verify that I want to make sure that they are exactly where they said they are. They're pretty close. Good job. But what we really get to know now is, ah, you didn't tell me about this rack up here. Good to know. And you can discover all of the features that might have been left out from someone's construction model or all of the field conditions that were encountered in the day two work that got changed from a design model that you got handed and said this is exactly what I did. Well, maybe. You didn't tell me about that pipe down low that's going to block my installation for everything above it.

So, at this point, I'm happy to show off the software. And we can do some of these examples, and fly through them, and walk through them. Are there any questions first right now? Yes, sir.

AUDIENCE: Did you [INAUDIBLE]

JACOB RASKIN: There is a way, and I can try to show you. Maybe we'll work on that afterwards. It's a couple of steps that aren't quite as fast as I'd like to--

AUDIENCE: [INAUDIBLE] It's not something that [INAUDIBLE].

JACOB RASKIN: Yeah, it's not the most amazing workflow. Unfortunately, it's not, for example, quite as easy as Navis, where you can kind of point at the file and say relocate the file, kick it over, and respin it, and turn it upside down. And it records all of those features, and then spits them back to you. There is a way to do it in ReCap, though. It's a little tricky. And it's all about grabbing the features within the scan and saying, I want this wall to be facing north, rather than saying, I just know I want to take the whole scan and spin it 77.3 degrees. Any other questions? Yes, sir.

AUDIENCE: The portability of the spheres--

JACOB RASKIN: Yes.

AUDIENCE: So I got--

JACOB RASKIN: The sphere targets or the--

AUDIENCE: The sphere targets.

JACOB RASKIN: Yes.

AUDIENCE: So they're usually steel [INAUDIBLE]

JACOB RASKIN: That's what I have, too, yep.

AUDIENCE: So if [INAUDIBLE]

JACOB RASKIN: That's a great question. The question is, if you're traveling a far distance, for example-- getting on an airplane was this specific example-- how do you get around lugging along all of these spherical targets? It's the physical world holding you back from the digital world unfortunately. I'm really fortunate right now that I don't have to go very far. I haven't had to run up to that example.

AUDIENCE: [INAUDIBLE]

JACOB RASKIN: Yeah, it's not don't leave home. It's just bring your credit card, and go out to Michael's, and buy some spherical styrofoam at the local Michael's store. That's not a great workaround, and I've been told specifically don't do that. But I don't know. If you've got some acrylic paint that had a nice matte finish, I'd love to hear an example of I cheated, and it did work, even in Alaska. Yes, sir.

AUDIENCE: Just [INAUDIBLE] Do you have any best practices for that?

JACOB RASKIN: Sure. Can I repeat the question? The question is, I'm going out. I'm scanning my site. And the features in the ceiling, specifically these mechanical features, are dark and are not reading well or at all in my laser scan. What should I do? Is that the gist of it? OK. Are you scanning in color?

AUDIENCE: Yes.

JACOB RASKIN: OK. You might try scanning without color some time, or look to run it in that intensity view before you go out and send everybody out again to do the whole process. The guy who taught me actually encouraged me to not run in color. Our unit does it, but the few times that I have flipped the switch on, just to kind of do an example and say, let's do a comparison, I found that the white balance is actually somewhat tricky for the scanner to get correct on its own. And there's some photographic exposure settings generally with your scanner that you can toggle with.

But in my experience, running it in black and white has actually been pretty reliable to run in either a perfect wide open sky, sunny day, cloudy day, but also to go down into the bowels of Penn Station, where I spend most of my week, and scanning underneath this train bridge. And there's nothing but construction lights and a couple of sodium bulbs on the wall. So even though you say, OK, I don't want to give up that color. Maybe there's some other feature that was valuable to gain that color for. It might just be worth first reading in the intensity view to see if the points got read at all, or if they're just kind of being painted black because the exposure was off. If they really didn't get read, it might be valuable to go back out and try running it with no color and just trying to gain the black and white values. Yes, sir, in the back.