First they perform a rouch scan in market mode, which results in creating a marker point cloud.
Then they scan again, but this time with higher precision using the marker pointcloud for alignment - this way the high resolution scan doesn’t get lost as easily as without it.
Also other system have a similar two scan approach, when they first make a lo res scan from a higher distance, just to get a rough point cloud and shape understanding - then they make the proper scan but use the first rough shape to help in aligning the proper scan.
I wish , but I am afraid the scanner do not have the ability or proper sensors to do that , there would be needed another source of sensor to track the marker volume accuracy . Right now the sensors uses it for visual tracking purpose only .
But the Revoscan clearly recgnizes the markers during Marker mode. Why not just ignore any other data during the initial scan - the hard part is already done, the markers get recognized
The markers are recognized as a visual tracking stabilizer. Same thing when you add additional objects in Feature Mode to scan smaller volumes and that is not the hard part , the hard part would be to pre-build a volume with proper measurements based just on the markers and this is a next level stuff . I know exactly what you wish for, and I do too , who knows what Revopoint come up with next. Sooner or later they will need to evolve more to keep up with the 3D scanners business , but don’t expect POP2 to do that , there will be probably a new scanner and software to achieve this level and much more expensive.
The same with POP2 resolution , everyone wanted higher , well I tested higher resolution settings and the results was not good , that is when you reach the maximum the device sensors and algorithm can offer.
There are several techniques you can use to accomplish what you want. Do a fast scan with the Pop2, fuse and save. Then use the Mini to create smaller detailed scans. Use the fast scan to build your detailed scans upon. The fast scan is the skeleton upon which you build. Before doing final alignment remove the skeleton scan. Use global registration.
I use a lot of tools and fixtures when scanning, I do my best to maintain as close to ideal conditions to get repeatable results.
Jeff. , other programs allows you to do raw alignment first , you can always remove the base skeleton before final global registration ( the final fine alignment ) all the base alignments are not fused yet so you can add as many you want or remove , the global registration will fuse what was left.
Erked that’s not the point at all. The rough scan cloud is for the scanner to help it not get lost. If you’re going to stitch the scans afterwards, why do you need the “skeleton” scan - unless your detailed scans are not overlapping. So far I think the biggest flaw of the Revoscan is that it often gets lost while scanning and can’t align new scan data. This would be one way to help with that problem.
We do a similar thing when 3D tracking camera shots - you give the software a rough estimation of how the camera was moving - this helps it track the motion precisely, since now it knows which results comming from the algoruthm alone are obviously wrong.
Kaczorefx, there are usually several way to work around issues with scanning. Without knowing what is going to be scanned, resolution desired, and end use of the scan it’s difficult to guve helpful suggestions. Even knowing all the above, it can still be difficult. You say that this is a short coming of the scanner. I can’t think of another scanner in this price range that even competes with Revopoint. Please correct me if I’m wrong.
Oh, I am faaaar from calling it a short coming I love my POP2. The hardware is already there - what I’m suggesting and what is shown on the video is simply a software solution. Sure, it’s not our idea, but it is an idea, no reason why it can’t be copied. Doing a rough scan first and then rescanning with higher fidelity has been quite common in different scanning solutions - sure they are much more expensive, but I don’t really see a reason why Revo shouldn’t be able to replicate it. It’s already doing part of it aligning new scanned data to already scanned data each time we resume scanning. All that is lacking is a concept of the “skeleton” point cloud that has been created previously, that would help orient the currently scanned data, that’s it
I apologize if I came across the wrong way. I would also love to have this feature as well. I would like to be able to use my phone as a scanning monitor for the Mini. It would make scanning easier not having to shift my eyes far from what I’m scanning. I don’t need it to do any processing, although being able to save scans of point clouds from the phone would be great.
If you going to use MINI with your phone , you can keep using POP2 . Phones can’t process point cloud with the resolution at 0.02mm , not enough processing power and RAM .
But what you can actually do is link your computer screen to your phone and have high quality scans, monitoring your scanning via phone screen or better tablet but process it with the power and speed of your computer at maximum quality .
Already tried it, the video lag is terrible - not that it is very long, but it’s about half a second - it’s enough to interfere with your brain and I keep overcorrecting my scanning motions. So far the best I managed to do is mount my Surface tablet onto the Pop2 - or vice-versa, hard to tell really - you get a huge screen for realtime preview and you get the benefit of actually saving your unfused scan data, and fusing it later on. There is a new tablet coming to Kickstarter on june 9th, it’s a Surface Pro rip off but with Intel 12th gen - this should actually be enough to fuse the points on the tablet in quite good times. I might be tempted to invest actually.
I use TeamViewer , there is no delay and even gain control works with touch where it does not on my Wacom Cintiq Monitor and PC
You install one on your desktop and one on your mobile , the desktop will give you new password every couple minutes to use , easy and simple . You can zoom in and out for finer details as you want , I just made screenshots for you on my phone .
you can hide the menu bar .
I love my POP2. The hardware is already there - what I’m suggesting and what is shown on the video is simply a software solution. Sure, it’s not our idea, but it is an idea, no reason why it can’t be copied. Doing a rough scan first and then rescanning with higher fidelity has been quite common in different scanning solutions - sure they are much more expensive, but I don’t really see a reason why Revo shouldn’t be able to replicate it. It’s already doing part of it aligning new scanned data to already scanned data each time we resume scanning. All that is lacking is a concept of the “skeleton” point cloud that has been created previously, that would help orient the currently scanned data, that’s it 
