Creaform MAXshot Demo
While out here on a sales trip, Jarrod Schmidt – my Creaform sales rep, dropped by to demo the MAXshot option for my EXAscan 3D laser scanner. What is the MAXshot? In a nutshell, it’s a range extending option for the EXAscan. As I outlined in my first post on the EXAscan, the error is a function of both the laser (very small in this case) and the distance over which the scan takes place. The first term is fixed, and the second of course grows as the size of your scan grows. It’s called the volumetric accuracy of the unit. The MAXshot significantly decreases the error associated with the volumetric accuracy – by a factor of 4. So, for a scan with a 40′ diameter using the MAXshot, the error would the same as for a 10′ diameter scan with just the EXAscan – that is 0.0072″. Considering that’s the spec accuracy of my FARO mechanical arm, which had a reach of a mere 4′, that’s just astounding. So how does it work? Well, it adds a photogrammetry approach to the process. You start by placing the targets on your surface. With the EXAscan, there’s just one type of target – the small reflective circles. With the MAXshot add on, there are three additional target types. The first is what looks like a carpenter’s square, and is the first target to get captured. The next type of target is the coded square targets. They are all about 2″X2″ with unique dots on each one. There are hundreds of these, and the computer can tell exactly which one you are photographing. Lastly, there are scale bars, which are a known size. They are made of carbon fiber tube, for low thermal expansion, and have coded square targets on each end. Ideally, you fit two of these in your scan, oriented in different directions. In the picture below, you can see all four types of targets:
The overarching idea is this – the coded targets, reference bars and carpenter’s square all are used by the software to build a highly accuracy model of the spatial relationship between ALL of the targets. Essentially what you are doing is building a bridge between all the areas that you want to laser scan – the areas that have the regular circle targets. Once you have this photogrammetry model, you can take all the MAXshot targets off your surface, and then go about laser scanning your individual regions. Since the MAXshot sees the circle targets when you build the photogrammetry model, once you start laser scanning the computer can deduce where on your model you are scanning, and place that data in space accordingly. In order to build that model, you use this small camera supplied by Creaform, which is attached to the computer via firewire cable. You start by taking a 3X3 grid of photos of the carpenter’s square, and then move out from there. It’s important that the targets do not move relative to the target object during the acquisition, but once the photogrammetry model is complete, that no longer applies.
One thing that’s really neat is that you can still add targets as you go, after you have built the photogrammetry model. So, if you want to say scan the entire cowling, instead of just the nose bowl, you would just add the circle targets to the rest of the cowling and scan that area as well. What I really like about this setup is that you can really pick and choose just how much data you want. For most aircraft – especially those that are sheet metal – there are surprisingly few areas that need to be scanned to build a very accurate model. The wing, for example, might only need to be scanned in a few places. It’s really only the areas of compound curvature – nose cones, fairings etc – that need to be fully captured. The rest can be easily interpolated from section scans that need only be a few inches wide. Everything else is just useless data.
For this particular demo, we decided to laser scan three sections of the aircraft – the first of course being the nose bowl. We also added targets around the door latch on the right side of the aircraft:
Finally we added some targets around the air inlet on the left side of the aircraft, ahead of the pilot’s door:
The photogrammetry model for the entire aircraft looks like this:
It’s a bit hard to tell, but you can make out the cowling on the right of the screen. Just to the left of the cowling, by the target labeled 267, you can see the cluster of targets around the air inlet. On the left are the targets around the door lock. Once we laser scanned the local areas, we had this:
The purpose of the demo was the demonstrate the technology, not to capture every last tiny detail of the surfaces, so we kept the resolution quite coarse. Even still, you can make out some of the rivet details and such:
Given that I don’t have an immdiate application for the MAXshot I’m going to hold off on purchasing it for now, but it’s nice to know that if and when I need to extend my range, that the solution is already there. For the type of work that I do, I really like this approach to dealing with large scans. To be able to pick and choose what data you want from the outset will be a huge help in not getting a bunch of extraneous data, and as I outlined on my post about the coming T-Splines 3.3 release, taking this data and then making a very nice T-Splines/NURBS surface on top of it has never been easier or better.