Scanning of an action figure (Hawks from My Hero Academia) with dimensions 220x220x95.
The scanning was done in several stages: first, a partial scan of the entire model, and then scans of the torso, legs, and the torso with the wings.
The wings were later isolated from the model to have only the wings with the correct positioning.
Apart from the partial scan of the torso, which was done in a single phase, the subsequent scans were carried out in three phases, moving the scanner or rotating the model.
All scans were performed on a turntable with a rotation speed of 200s.
The point clouds were fused in Advanced mode at 0.3mm.
The partial scans, except for the wings, were merged in the latest version of RevoStudio5 (126.96.36.1992).
The wings were merged with the model manually in Revo Studio 188.8.131.52, which allows convenient fusion of partial parts by manual placement. However, the same operation can easily be performed in MeshMixer.
The generated file was then imported back into RevoStudio5 for mesh generation.
Subsequently, the mesh was oriented (straightened) in PrusaSlicer and is ready for use.
The obtained model is of excellent quality despite the complexity of the original model, managing to capture all the smallest details of the subject, from the jacket buttons to the slightly raised texture of the belt and the texture of the boot soles!
I hope you find the work to your liking!
Thank you very much. I usually use MeshMixer to manually position non-overlapping scan parts, but every now and then I do it with the old Revo Studio as well. As for the trick with PrusaSlicer to orient the scans, Jermaul suggested it to me, and I must say it’s really convenient!
Absolutely true. It’s a pity, I find the manual positioning function very convenient, as I often need to merge scans that don’t overlap much to avoid the issue of surfaces being damaged by mismatched points. I hope they reintroduce this feature in RV5.
Another thing that could be improved, in my opinion, is the point selection for deletion. If you can’t select only the specific points in certain areas, you can’t remove them, leading to the need for more work in Meshmixer or similar tools.
That why the Advanced mode use in its algorithms.
I saw you have lots of points that did not belonged to the main scan , but was probably too far to be detected and removed … in cases like that I move to CC as the manual isolation or overlapping point detection not always works efficiently.
In case of more complex models it is nesesery to overscan some area to reach all the details .
But it seems that all models fused with setting of below 0.2mm have that issue , why you think we getting all the noises when fusing at 0.1 or below , overlapped points …
Duckling with Hearts.
Scan To Print direct.
Scanning of a figurine depicting a colorful duck with various engravings and colored writings on the body, measuring 70x50x60.
The scan was carried out in Standard Precision mode with feature tracking on a turntable at a rotation speed of 200s.
The scan was conducted in a single session with 4 repositionings of the scanner and the subject.
The point cloud was fused in Advanced mode at 0.2mm and the mesh was created at level 6.
The resulting file did not require any cleaning or optimization and was sent directly for printing.
The model is flawless, with the engravings captured excellently even though they are colored in black and are only a few tenths deep.
An example of straightforward Scan2Print.
You see just because you used Advanced mode at 0.2mm don’t means you lose accuracy at all , it just tell the algorithms to ignore the noises that you normally still have to remove when using Standard mode what is usually 50% of a junk and trash data …
I am using now Advanced mode for everything with a perfect results each time and great details , Standard mode don’t give you more details , it give you more noises and junk data …
Scanning of a battery-powered screwdriver with dimensions of 220x160x80.
The scanning was done in Precision Standard mode on a turntable with a rotation speed of 200s.
6 repositionings of the subject and the scanner were performed.
The point cloud was fused in Advanced mode at 0.2mm, and the mesh was created at level 6.
The model turned out to be highly detailed and clean, requiring no post-production work. The details on the rubber and engraved plastics are fantastic, including the screen printing and even the screws inside the holes.
The model, necessary for creating some supports requested by a friend, was straightened in PrusaSlicer, decimated, and imported into Fusion360 to create the parts.
Once printed, the supports turned out perfect in terms of size and fits.
Already having the model of the screwdriver, I created a simple miniature keychain by just adding a ring to the screwdriver’s head… couldn’t be easier.
What do you think?
You’re absolutely right. I’m finding the advanced mode really convenient, and especially when looking at the latest scans of the screwdrivers, you can see that the detail is really high, and there’s no loss in terms of quality! It also makes the process much faster because it almost eliminates the need for post-production on most jobs!
Cloth Owl Piggy Bank.
Scan of a cloth piggy bank in the shape of an owl, measuring 200x180x90.
The scan was performed in Precision Standard mode with feature tracking on a turntable at a rotation speed of 200s.
Depth camera exposure was left in automatic mode, while the RGB camera exposure was set to 46.
The scanning was done in a single session with 6 repositionings of the subject and the scanner.
The point cloud was fused in Advanced mode at 0.36mm, and the mesh was generated at level 5. The result is excellent, the mesh is clean and captures the details of the cloth and its texture perfectly.
What do you think?
Rononoa Zoro Action Figure from One Piece.
Given the tremendous success of the live-action adaptation of Eiichirō Oda’s manga One Piece, produced by Netflix, I couldn’t resist paying my tribute with a scan of Zoro’s action figure, one of the main characters of the series.
The action figure measures 300x220x180.
The scanning was done in several stages, all carried out in Precision Standard mode with feature tracking on a turntable rotating at 200s.
The subject was initially scanned in its entirety, with multiple repositions of both the subject and the scanner. Subsequently, details such as the swords were scanned separately and then merged with the original model using RevoStudio5 for the fusion of various body parts, while RevoStudio (184.108.40.206) was used for the manual positioning of the swords.
The resulting model is highly detailed and clean, with excellent quality in small parts and incredible anatomical detail in the muscles!
The scanner’s ability to capture hard-to-reach points on the underside of the model is also remarkable.
I hope you find this work enjoyable.
What do you think?
Thank you very much, Fabio.
The images you’re referring to are part of a reverse engineering process.
Essentially, I import the scanned image into Fusion360, the one composed of numerous faces, and use it as a reference to redraw a clean and editable CAD model.
In Fusion360, but only with the commercial license, there is a function to convert from Mesh to Solid, but honestly, it doesn’t work well with scanned images, partly because 3D scanning-derived meshes have too many faces to be processed by the program.
In reality, even with ‘clean’ meshes from previous CAD designs, the conversion function doesn’t work perfectly, except for very simple and linear geometries.
there is a method that results in better geometry specially for clean or organic surfaces that can even be reversed with surfaces after on free fusion360, that is to use the opensource instant meshes software to convert the mesh to clean quads.
I have used that method for the reverse engineering of a hourglass here