as promised I will post the result of the scan of a 72 cm tall sculpture, completely made with the
MINI. I made a series of scans (0.02 resolution) which I then merged into Meshlab and meshed with CloudCompare. The model obtained rather large (over 4 million points) but the result satisfies me.
As you can see the porous, eroded, wrinkled surface has been acquired by the scanner. The fine details of the surface modeling are recorded in the 3D model.
Below is the result:
Beautiful work Davide ! the reliefs in the back are so much sharper now too . Well done !
thanks for sharing ! must take you a while to put it all together ! and probably lots of data at that size …
Very careful to move the scanner smoothly and “some” alignment work
Big scan data (files from 500 mb to 1,5 gb)
Well done. Thanks for sharing. I am sure that your showcase with data will help others.
Thanks for appreciation
Really amazing what these things can do in the hands of people who know how to get the most out of them! Thanx for sharing!!!
Thanks for your enthusiastic appreciation
Excellent post, the details are excellent comparing this scan of the sculpture to your POP1 post.
Did you find using the MINI freehand easy to use? I have seen others mention the MINI is more sensitive to changes in the distance from the object during scanning. Thus, the scanner mounted on the stand is better as the distance is the same the whole time (with the object on the turn table). But of course for free hand scanning this is not possible.
Also, did you find tracking to be good for this large object? Approximately how long did it take you to scan the object (just point cloud)?
I use MINI with freehand for most scans I make (because the works I scan are not positionable on the turntable, both in size and weight).
You have to have your hand still but you can do it. The use of the gimbal for the MINI will be a great help.
The scan was done in several portions aligned in MeshLab and the point cloud was meshed in CloudCompare. For the scan the time it took was half an hour. Fusing the point cloud is the one that takes the most time (at resolution of 0.02).
It all took me about 2 hours away (scan, fusing the point cloud, editing and aligning in Meshlab, meshing in CloudCompare).
What happen when you move the scanner too far or too fast? does it loose instantly the tracking point? does the software go back to the previous point where it had succesfull data adquisition? would you have to repeat the whole recording from the space point where you started or simply from where you lost track?
Sorry about all the questions but i still have many doubts about how the scanner works in those situations.
Thanks in advance
The strategy for a quality scan is: uniformity, uniform distance, uniform speed.
If it loses the track it manages to resume tracking, if has enough points.
Thank you for your reply and the additional information regarding processing times. I look forward to your future posts.
So the handheld stabilizer won’t help with speed, neither with distance. That’s your own skill and practice. If any, it will help to maintain the angle/ tilt towards the object (hoping it has the ability to block itself at a giving angle).
@PUTV suggested that there was an optimal orientation angle to scan with MINI (probably also with others, i don’t recall exactly). How crucial is that inclination? is it related to the quality of the point cloud or more about the ability to maintain tracking?
@J0sh the tracking will depends of the feature of the scanned object , MINI can pick up more details for that reason it is more stable while tracking .
The speed as always depends of how many frames the scanner can capture per second , how more frames it can capture per second how faster you can move while scanning .
All of it depends on what you are scanning so there is no one answer for everything .
If you go too slow you will get overlapped frames, if you go too fast you will get vertical strips on your model surface.
There is also the angle , at 45 degrees it captures the details well but that is because at that angle it can see more faces of the object than on other angles .
And stabilizer… can’t say much about but it
because none of us tested it yet, but it has to have tilt at angles or you would be not able to scan anything with .
And it will definitely help capturing better frames for the point cloud . I don’t see how it could helps with tracking, as tracking are based on visual features of the object no matter your hand is stable or not .
How many scans did you do for this whole model?
Hi Jorja Albert,