Scaning shining and metal artifacts


(Nieda) #1

Hello guys,
I've came across a difficult task: scanning a bronze shining art sculpture which is part of a very important (and expensive) brazilian museum collection.
After 3 attempts, we managed to scan it properly.
To get rid of reflections, we used a wax to cover the surface, which was applied tapping the surface, which left a very tiny texture.
As for the camera equip, I used a Canon 70D (raw) with canon 10-18mm STM (@10mm).
To keep the F stop at its smallest I used 3 flash heads with 100cm softboxes.
I was able to create the model with Memento.

Any other ideas guys? which method are you guys using in cases alike?

The model:

Cheers,
Vitor Damiani


(Abby Crawford) #2

Great model, @nieda! I don't have experience with shiny metals myself, but feel like I've heard of people using a really fine dust to limit shininess and give objects textural points. Do any of you have experience with this, @micropasts, @nebulousflynn, @barbierilow, @bdnewnan, or others who I haven't mentioned?


(Nestor F. Marques) #3

Hi @nieda, your solution is great! The wax worked perfectly!

I have some experience with shiny, metal and reflective objects. I have to say that I've tried various methods to do it. I'll show you some examples. I usually work with archaeological and historic artefacts that can't be rubbed with wax or dust (I wish I could...)

This is a very reflective silver coin due to it iridiscent patina. I managed to scan it perfectly with a light box with very soft light in photoscan.

Cincuentín de 1620 Felipe III MAN by Néstor F. Marqués on Sketchfab

This greek vase was also quite a challenge at the time, it was not perfect, but I managed to scan it in day light (cloudy day light to be precise)

Cratera griega Museo Arqueológico Nacional by Néstor F. Marqués on Sketchfab

And finally this is one of my favourites because it's a glazed pottery plate with metallic reflex. It was achieved in studio conditions with 100cm continous light softboxes, turning the object and masking each photo separately.

Plato persa / Persian plate by Néstor F. Marqués on Sketchfab


(Nebulousflynn) #4

I agree - really nice scan @nieda, and thanks for sharing your process.

I've not had much luck myself with shiny objects, although this brass door knocker came out well, perhaps as it had a lot of non-shiny context around it? https://skfb.ly/C6Bp

At the moment it seems that light based scanning would indeed require some re-texturing of a shiny surface - @3digify use an anti-glare spray for example: https://skfb.ly/EqoY - useful if the object is not precious or ancient but I'm not sure curators / collectors would allow such contact with their objects.

I've heard of "contact scanning" which uses a robotic arm to record points in 3D space too, but I guess you don't get a textured model at the end of it...


(Nebulousflynn) #5

Update on this, I made a quick scan of a gold mask using a small portable light tent in an 'office lighting environment" : https://twitter.com/_museuminabox/status/664072182612537345

It's only one sided but a good outcome for the test. I used a Photoshop action to auto-magically create masks by selecting the white background. What do you think!?

Gold Mask by nebulousflynn on Sketchfab


(Nestor F. Marques) #6

Ah, nice scan @nebulousflynn! Now I want to see it from the back!

I've made some improvements myself regarding shinny objects. Here is a 1.3 meters tall apulian lutrophoros. Very shinny surface.

Lutróforo del rapto de Perséfone by Néstor F. Marqués on Sketchfab


(Nebulousflynn) #7

Beautiful! Did you use any special technique, @NestorMarques ?


(Nestor F. Marques) #8

Hi again @nebulousflynn, thanks and sorry for the delay!

It was made under studio lighting conditions, which are very good for this matter. As I've seen in your twitter, you did too.


(Lapin) #9

3-D Laserscanning Antiglare Spray (Helling, Germany). Or any other equivalent.


(Ben Kreunen) #10

I came across a spray used in conservation inspection processes that dissipates after a while but unfortunately I forget the name of it ... will ask our conservation department. It's primary function wasn't to reduce reflections, but it was a handy characteristic with the benefit of being temporary. i haven't done many shiny objects, but one strategy I employ if to take many more photographs at small increments. This allows for better local detail as the reflections don't change the appearance of the surface as much


(Abby Crawford) #11

@port0228 what product do you use to cut down on the shine of lithics? Would it be suitable for ceramics and other more delicate objects?


(Samantha Porter) #12

I'm actually currently working on this issue for a presentation at the upcoming CAA conference in Oslo! Specifically, we're comparing different coatings as well as different scanning methods (i.e. structured light vs. photogrammetry) for use with obsidian artifacts, which are frustratingly both shiny and clear.

Normally, for lithics I use developer spray. This is what I used to coat my core models from Les Cottés, and what my colleagues and I published in our recent paper in Advances in Archaeologial Practice (you can download the article from my website here). I would definitely not advocate using this on ceramics, or anything that is porous. It needs to be cleaned off with water. Furthermore, it contains some acetone. Again, not a big deal for stone tools (although we commonly do have to re-label artifacts after use) but possibly a big problem for most other things.

Recently, we've had some good luck with dulling spray (see model below). However, this stuff only works with structured light scanning, and not photogrammetry. We also wouldn't really recommend using any of these coatings on objects that may have residues, or that will be subject to any type of chemical analysis.

Experimental Obsidian Point by Samantha Porter on Sketchfab

The next thing we'd like to try is cross-polarized photography. Anyone here have any experience with that?


(Ben Kreunen) #13

Depending on the material, and if geometry only is required I've referred a few people at work to the 7T MRI scanner at our uni. Unfortunately didn't work for "potted" specimens in formalin.

I've seen a few samples using cross polarisation and it works well if you have the flash power. Used to use it on our copystand for all reprography.


Share Gear / Software or Hardware Comparisons / Scan Experiments
(Nebulousflynn) #14

@NestorMarques @lapin @uomdigitisation thanks for tips! @port0228 - what a great model, very cool :sparkles:


(Nebulousflynn) #15

@otto_bagi just shared this video of @port0228's CAA presentation over on this thread, thought it best to share here too - very interesting and thorough research!


Add new "3D scanning and photogrammetry" group in forum
(Otto Bagi) #16

I was experimenting with the team in the Institute with cross polarisation and I must say it works pretty well.

This setup, in effect, cuts off almost all surface reflections, although, it requires a bit of work with the shader on sketchfab to get back the original appearance of the object.

Also, the photographing can be time consuming but this depends on the strength of the light you use. Since the polarising sheets cut off a fair amount of the light, in this particular case we had to use 8-10 sec shutter speed to compensate, but I think the final result had made that 5 sweaty hours in the studio in 35C worthwile. :slight_smile:


(Lapin) #17

James Busby wrote a detailed tutorial on scanning reflective surfaces and shared with everyone. At the end of the tutorial there is a link to download the scene of the lesson.
A big thank you James!

3D Scanning Reflective Objects With Photogrammetry
By James Busby


(Abby Crawford) #18

@lapin that's a great tutorial! Thanks for passing it along. As the author notes in his intro, it's not likely to help with expensive or historical objects that can't be covered in matte-ing (matting?) agents, but it's an extremely useful tutorial nonetheless. :slight_smile:


(Mauricesvay) #19

I've heard that some people use talc to make shiny objects less shiny. I haven't tried though.


(Fredfroehlich) #20

@uomdigitisation I guess the name is Cyclododecane.