3D scanning - high resolution for small objects

Write here about construction, ideas, equipment, tips n tricks etc. related to structured light scanning
barryvanbakel
Posts: 10
Joined: 02 Jan 2021, 14:58

Re: 3D scanning - high resolution for small objects

Post by barryvanbakel »

So if I understand right;
- The projector should be as close to the subject as it can still get in focus, and
- The cams should have the subject as frame-filling as possible.
Correct?

Then, does it matter how tumid an object is? A coin is flat subject, OK, but when projecting a pattern on a semi-globular object, lets say a walnut, then do both cams record the pattern correct? And I imaging the projector does have a focus plane so with a tumid subject there will be parts of the projected pattern out-of-focus?
User avatar
Micr0
Posts: 514
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: 3D scanning - high resolution for small objects

Post by Micr0 »

barryvanbakel wrote: 07 Jan 2021, 13:21 So if I understand right;
- The projector should be as close to the subject as it can still get in focus, and
- The cams should have the subject as frame-filling as possible.
Correct?

Then, does it matter how tumid an object is? A coin is flat subject, OK, but when projecting a pattern on a semi-globular object, lets say a walnut, then do both cams record the pattern correct? And I imaging the projector does have a focus plane so with a tumid subject there will be parts of the projected pattern out-of-focus?
Yes and no. there is a sweet spot where the scanner is most accurate. That does have to do somewhat with the focus of the projector (more the projector than the cameras at small apertures). If you have an object that you have set up correctly and used the correct calibration scale it should be fine.
µ
User avatar
OBNRacerMan
Posts: 287
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: 3D scanning - high resolution for small objects

Post by OBNRacerMan »

I would like to add that the closer the subject is to the lens, the shallower the depth of field is obtained (at least taking into account the aperture, at least without).
This was one of my options
This was one of my options
Therefore, the best option for macroscanning is to reduce the projection size of the projector to the desired one (by replacing the lens on it, or by choosing the lenses), but so that it is from a distance of half a meter to a meter. And use lenses for cameras with a focal length of 35-55 or even 75mm.
In this case, we get a compromise between a sufficiently large depth of field and susceptibility to vibration.
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: Byintek UFO R9, Cam: Daheng Mercury MER-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC HM1628MP10 (F2.8, 10Mp, 16mm) x2 & Azure 3518M3M (F1.8, 3Mp, 35mm).
barryvanbakel
Posts: 10
Joined: 02 Jan 2021, 14:58

Re: 3D scanning - high resolution for small objects

Post by barryvanbakel »

I see, more distance is more DOF, and longer focal lengths needed.

Projectors with replaceable lenses are expensive > is there an option to 'stack' a lens in front of the projector lens; or, to put an attachment lens onto it, in order to get a different angle of the projector light beam?
User avatar
OBNRacerMan
Posts: 287
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: 3D scanning - high resolution for small objects

Post by OBNRacerMan »

Interchangeable lens projectors are very expensive (starting, probably, from 4-5 thousand dollars ...)
You can try to remedy the situation with additional lenses - but for this you need to have a whole set to determine which magnification / reduction value is most suitable ... unfortunately, I do not have this opportunity.
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: Byintek UFO R9, Cam: Daheng Mercury MER-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC HM1628MP10 (F2.8, 10Mp, 16mm) x2 & Azure 3518M3M (F1.8, 3Mp, 35mm).
User avatar
Micr0
Posts: 514
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: 3D scanning - high resolution for small objects

Post by Micr0 »

OBNRacerMan wrote: 08 Jan 2021, 19:43 Interchangeable lens projectors are very expensive (starting, probably, from 4-5 thousand dollars ...)
You can try to remedy the situation with additional lenses - but for this you need to have a whole set to determine which magnification / reduction value is most suitable ... unfortunately, I do not have this opportunity.
The problem with add on lenses that change the focal length is that they are usually very cheaply made (home theater) and are very distorting. I'v tried a number of even name brand (tiffen) and have never had them work well. Even if I could get David to calibrate using one of these, the scans were usually unusable.
µ
User avatar
OBNRacerMan
Posts: 287
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: 3D scanning - high resolution for small objects

Post by OBNRacerMan »

Micr0 wrote: 08 Jan 2021, 20:33 Even if I could get David to calibrate using one of these, the scans were usually unusable.
It's easier for me in this regard - Rangevision does not use a projector during calibration, so I scanned with a lens in front of the projector with the same result as without it. But that lens (ordinary magnifying glass) that I have did not allow me to reduce the projection field as much as I would like, although the pixel density per unit surface area increased slightly.
Comparing results adding lens to projector
Comparing results adding lens to projector
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: Byintek UFO R9, Cam: Daheng Mercury MER-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC HM1628MP10 (F2.8, 10Mp, 16mm) x2 & Azure 3518M3M (F1.8, 3Mp, 35mm).
barryvanbakel
Posts: 10
Joined: 02 Jan 2021, 14:58

Re: 3D scanning - high resolution for small objects

Post by barryvanbakel »

I suppose to use an enlarging lens in front of the projector could be worth trying, they are designed to provide minimal distortion. And affordable.
But perhaps it is not necessary.... the coin of the first reply shows remarkable detail allready.
Does anyone know if the AAXA P7 or P5 are tested in a structured light scanning setup?

Kind regards,
Barry
User avatar
OBNRacerMan
Posts: 287
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: 3D scanning - high resolution for small objects

Post by OBNRacerMan »

My HD projector (1280 * 720) has a throw ratio of 1: 1.47 with a focusing of 100mm, AAXA projectors have a TR 1: 1.2 - i.e. if they can focus from the same distance, they will produce a larger picture, but due to the fact that there is a FullHD pixel from the projection onto the object, it should turn out to be about the same size (but if the projectors focus only from a greater distance, then the FullHD pixel will be larger than HD at close range). The larger the pixel in the projection, the worse the quality of the resulting surface.
Soft: RangeVision DIY, Calibration panel/Rotary tables - RangeVision, Prj: Byintek UFO R9, Cam: Daheng Mercury MER-630-60U3M-L (USB3.0, monochrome) x2, Obj: ZLKC HM1628MP10 (F2.8, 10Mp, 16mm) x2 & Azure 3518M3M (F1.8, 3Mp, 35mm).
User avatar
Micr0
Posts: 514
Joined: 15 Nov 2016, 15:20
Location: New York City

Re: 3D scanning - high resolution for small objects

Post by Micr0 »

OBNRacerMan wrote: 10 Jan 2021, 08:42 My HD projector (1280 * 720) has a throw ratio of 1: 1.47 with a focusing of 100mm, AAXA projectors have a TR 1: 1.2 - i.e. if they can focus from the same distance, they will produce a larger picture, but due to the fact that there is a FullHD pixel from the projection onto the object, it should turn out to be about the same size (but if the projectors focus only from a greater distance, then the FullHD pixel will be larger than HD at close range). The larger the pixel in the projection, the worse the quality of the resulting surface.
I guess we will all have to upgrade to 4K projectors. :)
µ
Post Reply