First scans with lowbudget rig

What have you scanned recently?
User avatar
Micr0
Posts: 427
Joined: 15 Nov 2016, 15:20

Re: First scans with lowbudget rig

Post by Micr0 » 27 Aug 2020, 13:50

Tzk wrote:
27 Aug 2020, 11:22
OBNRacerMan wrote:
03 Aug 2020, 16:21
My RMS value is more affected by the resolution of cameras (if you make a low resolution, then the result will be better on the same table) I also noticed that if the table is smaller than the area indicated in the center of the program screen, then the edges of the scan, even with a good RMS bend strongly

I noticed that a) my scans bend and b) RMS gets insanely bad when i'm too far away from the panel. Here's two calibrations with RMS 0.2 and 0.4 for comparison. I only moved the panels about 20cm backwards and adjusted the focus for the 2nd calibration.

I guess that's how i got RMS 0.13 when i first tried the setup... Anyways, i'll keep an eye on it in the future.

Capture.JPG

Capture2.JPG

I also practiced a bit. Here's my trusty Makita drill. I had to take ~70 scans and threw 30 away (so 40 used for the fused result). I guess that are too many scans, right? I also only scanned one half of the drill...

Capture3.JPG
I didn't think David was that sensitive to the distance of the cameras to the panels. I have always tried to have my scan subject more or less in between the outer most donuts. I Have a piece of string tied to the mount for my projector that has small knots tied like graduations so that I can place the items I'm scanning exactly in the sweet spot where the calibration panels were. I do remember seeing a hysteresis map of the error in scan data vs the distance from the focal point of calibration with the calibration corner. It drops of fairly rapidly after you are 2 or more "dots" from the center donuts. Better to use a properly sized calibration panel than to try to scan something larger than the calibration should allow.
µ

Tzk
Posts: 23
Joined: 17 Mar 2020, 13:28

Re: First scans with lowbudget rig

Post by Tzk » 27 Aug 2020, 16:02

I'd say it's not the distance which caused these high RMS values but rather the wider view onto the panel. If i barely fit the donuts into the view, then RMS drops. If i include the whole panel (and thus the donuts become smaller) RMS increases. So your conclusion to a) use panels which are big enough for the object and b) to place the object between the donuts is spot on.

I'm curious: How do you approach really big objects like a car? I guess i need huge panels for this and then scan from further away to get the general shape and then again for the details, possibly with smaller panels? I've done a car with photogrammetry in the past but not with SLS, that's up next when i'm familiar with the scanning technique.

User avatar
OBNRacerMan
Posts: 176
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: First scans with lowbudget rig

Post by OBNRacerMan » 27 Aug 2020, 19:51

In the case of David / HP3DScan software, only a huge table will help (because this program does not support markers).
I myself am now switching to Rangevision software as it has support for markers, which allows using relatively small tables (square with a side of 300mm) even for large objects (and, besides, this program is actively developing).

P.S. For these reasons, I sold both my SUA630M cameras (the new software does not support them). I also sold good 5MP lenses ZLKC 12mm F 1: 1.4 with cameras.
New cameras (essentially the same as they were, but from a different manufacturer and with different drivers) will come to me soon. I also bought a new projector (now with FullHD resolution) and new 10MP 16mm F 1: 2.8 lenses (but later I will also buy more lenses with focal lengths of 12mm and 25mm - for scanning large and very small objects).

User avatar
Micr0
Posts: 427
Joined: 15 Nov 2016, 15:20

Re: First scans with lowbudget rig

Post by Micr0 » 28 Aug 2020, 14:22

OBNRacerMan wrote:
27 Aug 2020, 19:51
In the case of David / HP3DScan software, only a huge table will help (because this program does not support markers).
I myself am now switching to Rangevision software as it has support for markers, which allows using relatively small tables (square with a side of 300mm) even for large objects (and, besides, this program is actively developing).

P.S. For these reasons, I sold both my SUA630M cameras (the new software does not support them). I also sold good 5MP lenses ZLKC 12mm F 1: 1.4 with cameras.
New cameras (essentially the same as they were, but from a different manufacturer and with different drivers) will come to me soon. I also bought a new projector (now with FullHD resolution) and new 10MP 16mm F 1: 2.8 lenses (but later I will also buy more lenses with focal lengths of 12mm and 25mm - for scanning large and very small objects).
I just looked into the Reangevision software. It look interesting. No prices on the web site though. Did you get a quote for just the software? if so how much is it?
µ

User avatar
OBNRacerMan
Posts: 176
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: First scans with lowbudget rig

Post by OBNRacerMan » 28 Aug 2020, 18:45

Micr0 wrote:
28 Aug 2020, 14:22
I just looked into the Reangevision software. It look interesting. No prices on the web site though. Did you get a quote for just the software? if so how much is it?
I am sure that at the initial stage (for Russian amateurs who expressed a desire to buy their software), the price was made lower (about USD350) than it will be soon after the "RangeVision DIY 3D scanner" program was put into operation.
Therefore, from September the terms of delivery will be slightly changed (for newly purchased ones, the key will have resolution restrictions (so as not to create direct competition with their flagship model Rangevision Pro, which also uses 6.3MP cameras, but produced by IDS, not Daheng)).
Over time, this should likely become available to overseas buyers as well (currently, resellers only sell original scanner kits).
The kit itself includes a program, a licensed USB key, calibration tables for three different zones (you can also additionally order an automatically rotary table controlled from the program). Cameras and lenses sold separately (owner).
By decision of the management, they limited themselves to using cameras only by IDS, Dahend and Toupcam.
You can try to find out more subtle in the Telegram channel - RangeVision DIY 3D scanner

User avatar
Micr0
Posts: 427
Joined: 15 Nov 2016, 15:20

Re: First scans with lowbudget rig

Post by Micr0 » 29 Aug 2020, 13:16

OBNRacerMan wrote:
28 Aug 2020, 18:45
Micr0 wrote:
28 Aug 2020, 14:22
I just looked into the Reangevision software. It look interesting. No prices on the web site though. Did you get a quote for just the software? if so how much is it?
I am sure that at the initial stage (for Russian amateurs who expressed a desire to buy their software), the price was made lower (about USD350) than it will be soon after the "RangeVision DIY 3D scanner" program was put into operation.
Therefore, from September the terms of delivery will be slightly changed (for newly purchased ones, the key will have resolution restrictions (so as not to create direct competition with their flagship model Rangevision Pro, which also uses 6.3MP cameras, but produced by IDS, not Daheng)).
Over time, this should likely become available to overseas buyers as well (currently, resellers only sell original scanner kits).
The kit itself includes a program, a licensed USB key, calibration tables for three different zones (you can also additionally order an automatically rotary table controlled from the program). Cameras and lenses sold separately (owner).
By decision of the management, they limited themselves to using cameras only by IDS, Dahend and Toupcam.
You can try to find out more subtle in the Telegram channel - RangeVision DIY 3D scanner
I was hoping it would be a successor to David. I' have been doing well with David, but after HP bought them it was essentially over. No upgrades etc.

The David calibration process is what bothers me the most about David. It seams you have to be overly careful to get really good results. I think David did it to make the process faster, but it is too prone to errors.
µ

User avatar
OBNRacerMan
Posts: 176
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: First scans with lowbudget rig

Post by OBNRacerMan » 29 Aug 2020, 14:33

Micr0 wrote:
29 Aug 2020, 13:16
I was hoping it would be a successor to David. I' have been doing well with David, but after HP bought them it was essentially over. No upgrades etc.

The David calibration process is what bothers me the most about David. It seams you have to be overly careful to get really good results. I think David did it to make the process faster, but it is too prone to errors.
Yes, it has been almost two years since the last version of HP3DScan 5.60 was released
In contrast to this, Scan Center software from Rangevision is actively developing, it has feedback from users, according to the wishes of which (if there are good reasons) additional functions are introduced. And now they have officially decided to pay attention to home-made people, providing an opportunity to buy software at an affordable price. And, what is very important, this software supports markers that allow you to capture large objects with a very modest size calibration tables.
Calibration in RV is more difficult, you need to take about 11 pictures of the calibration table from different angles. And here the projector is not involved in calibration. But for the same reason, this program requires the presence of 2 cameras (the same Flexsan3D has the ability to use 1 or 2 cameras).

Tzk
Posts: 23
Joined: 17 Mar 2020, 13:28

Re: First scans with lowbudget rig

Post by Tzk » 30 Aug 2020, 08:45

I've visited a local store and let them print a DIN A0 sheet for a big calibration table and will glue that to 19mm melamine-coated chipboard. The scale is ~420mm, let's see how that works out...

Rangevision sounds interesting. You've said that they introduced a resolution limit for diy scanners. Is it at 5mpix or even lower? What about camera support, are any UVD compatible cameras supported just like HP 3D Scan does?
OBNRacerMan wrote:
27 Aug 2020, 19:51
In the case of David / HP3DScan software, only a huge table will help (because this program does not support markers).
Can't we circumvent this by exporting the unaligned scans to a 3rd party software and align them there? However i'm not sure which software supports alignment by markers.

User avatar
OBNRacerMan
Posts: 176
Joined: 12 Aug 2019, 09:46
Location: Obninsk, Russia

Re: First scans with lowbudget rig

Post by OBNRacerMan » 30 Aug 2020, 11:28

Rangevision software supports only a limited number of camera models (the main ones are IDS and Toupcam), for DIY they made support for 6.3MP Daheng cameras (but limited them to 5MP, and in the future, for new buyers, they will be "cut" to 3MP), since the top model Their Rangevision Pro scanner also uses 6.3MP IDS cameras.
P.S. Regarding the use of scans in third-party programs - without markers it will be the same torment as in HP3DScan / David itself

Tzk
Posts: 23
Joined: 17 Mar 2020, 13:28

Re: First scans with lowbudget rig

Post by Tzk » 31 Aug 2020, 15:35

Having to use specific cameras and the limit of 3mpix doesn't sound that great. I'm still interested in your results with the new software.

----

I just finished building my 420mm panels and just placed them on the ground with some tape. The first alignment gave me a RMS value of 0.097 and i'm still very close to the panels. I'm a bit confused as micr0 talked about RMS 1 to 1.3 above, especially with big panels. So that confirms that RMS mostly depends on the size of the donuts on the image taken, i guess?
420mm.JPG
The panels are made from melamine chipboard and i used spray adhesive (or whatever pattex is called...) to glue the 100g paper to them. Works great so far.

Post Reply