Projectors... and Projector Mods.

Discuss about cameras, projectors, calibration panels, turntables etc.
erapip
Posts: 56
Joined: 16 Dec 2016, 07:16

Re: Projectors... and Projector Mods.

Post by erapip »

Hello bigbomber,

I find little time to test what you say. In attache you can see the results. How do they look?

Equipments:
1)Optoma EH 1020
2)uEye 1480-C
3) Original HP calibration board 120mm glass made
Attachments
2018-01-10_13-38-52.png
2018-01-10_13-38-22.png
2018-01-10_13-37-44.png
2018-01-10_13-21-29.png
bigbomber
Posts: 114
Joined: 14 Dec 2017, 18:08

Re: Projectors... and Projector Mods.

Post by bigbomber »

erapip wrote: 10 Jan 2018, 12:50 Hello bigbomber,

I find little time to test what you say. In attache you can see the results. How do they look?

Equipments:
1)Optoma EH 1020
2)uEye 1480-C
3) Original HP calibration board 120mm glass made
Thanks for your reply. You didn't mention calib scale.
Scan quality is ok. But there is alignment problem. As you can sed in last photo, some vertexes are off from circles center.
erapip
Posts: 56
Joined: 16 Dec 2016, 07:16

Re: Projectors... and Projector Mods.

Post by erapip »

It was a fast setup. I didnt optimize it. The cal. Scale is 120mm.
I never use two uEye camera (5mpx) because they give truble during datta capture ( one camera disconect). I think something related with DirectX. When only one is used everything is OK.
Should i test further? What about the Projector that i use, is it OK?

Thnx ;)
bigbomber
Posts: 114
Joined: 14 Dec 2017, 18:08

Re: Projectors... and Projector Mods.

Post by bigbomber »

erapip wrote: 10 Jan 2018, 18:39 It was a fast setup. I didnt optimize it. The cal. Scale is 120mm.
I never use two uEye camera (5mpx) because they give truble during datta capture ( one camera disconect). I think something related with DirectX. When only one is used everything is OK.
Should i test further? What about the Projector that i use, is it OK?
Thnx ;)
Actually this method I said (Align and compare a pair of single scan) is the best and absolute way to say how much your scanner is accurate.
via Dual cams, software create an average of scan between data captured from two angles. this reduce sensitivity to environment condition.
As a test: with single camera setup, create a single scan from an object and sent it to shape-fusion, and don't change its position in space . Without changing any things, just reduce projector brightness a little and create another scan and sent it to shape-fusion again. you can see scans are take distance from each other (like you change object distance from scanner setup). but with dual camera this error is reduced.

Can you calibrate your setup with single camera and do this comparison by scanning calib panel from two different angle (about 20 deg, but save Object-projector distance constant).

your projector seems not bad at a glance, but alignment of scans say something other.
I'm developing an application which can solve the majority of problem. it is in progress.

Edited Post: As I look more carefully on your last photo, I can see some jagged edges, this show your calibration is off, and a major reason for this problem (jagged edges) is not compatible projector. (I image your calib panel is perfect which is come from HP/David)
Capture.JPG
erapip
Posts: 56
Joined: 16 Dec 2016, 07:16

Re: Projectors... and Projector Mods.

Post by erapip »

Yes the cal. Pannel is supouse to be perfect.
I will try to better test. Can you explain better how to test with single camera? If the Total triangulation angle is 10deg, i have to test the second with 30deg (20deg difference)?
As i say i never use two uEye camera, they have problem with sincronization. Maybe this is the problem of jagged edges.
Thnx
mading
Posts: 307
Joined: 31 Jan 2017, 13:09

Re: Projectors... and Projector Mods.

Post by mading »

Asap I'll post some test. Thanks bigbomber for your help and interest.
I noticed a great loss in accuracy when using too small an angle (less than 20°) between projector and a single cam. The same 400mm long object showed 0.5mm and more deviations at its ends when scanned, flipped and scanned again. Think went better when I went>20°.
You can see examples of scanned panel in one of my previous post.
LG PF50, LG PF1500, RangeVision DIY: 2x DahengMer630, 2X12 and 2X16 mm 5Mp ZK lenses, RV turntable
bigbomber
Posts: 114
Joined: 14 Dec 2017, 18:08

Re: Projectors... and Projector Mods.

Post by bigbomber »

erapip wrote: 10 Jan 2018, 20:19 Yes the cal. Pannel is supouse to be perfect.
I will try to better test. Can you explain better how to test with single camera? If the Total triangulation angle is 10deg, i have to test the second with 30deg (20deg difference)?
As i say i never use two uEye camera, they have problem with sincronization. Maybe this is the problem of jagged edges.
Thnx
with single cam, just scan your panel simply. I mean just scan panel firstly from position1, then rotate it (about some degree, exp: 20 deg) and scan it again. now try to Fine-Registration these two scans, and create a comparison like before.
bigbomber
Posts: 114
Joined: 14 Dec 2017, 18:08

Re: Projectors... and Projector Mods.

Post by bigbomber »

mading wrote: 10 Jan 2018, 20:28 Asap I'll post some test. Thanks bigbomber for your help and interest.
I noticed a great loss in accuracy when using too small an angle (less than 20°) between projector and a single cam. The same 400mm long object showed 0.5mm and more deviations at its ends when scanned, flipped and scanned again. Think went better when I went>20°.
You can see examples of scanned panel in one of my previous post.
looking for your test photos.
in dual cam mode I use 18~20 deg between cameras, in both-setup mode.
erapip
Posts: 56
Joined: 16 Dec 2016, 07:16

Re: Projectors... and Projector Mods.

Post by erapip »

@bigbomber
I will try tomorrow some test. Also mine was 20 deg
Attachments
20180110_131849.jpg
mading
Posts: 307
Joined: 31 Jan 2017, 13:09

Re: Projectors... and Projector Mods.

Post by mading »

Well, I realised I have some room for improvements.. definitely more than expected! :lol:
(I'm competing for the most weird results)
A fast try (cams with different angles):
-wood 110 mm panel with photographic ink-jet printed paper glued with spray contact glue
-acer K132 (unmod)
-a couple of DMK 23UX236, at 1920*1080-60fps

calibration :
60.jpg
Various parameters:
prm.jpg
cam screenshots:
2018-01-11_12-11-13.png
2018-01-11_12-11-09.png
I got severe wave patterns, still looking for an improvement on it. Fine-tuning fps of projector and cams did not produced desidered improvements.

Results :shock: :
60fps-0.05.png
60fps-0.1.png
60fps-0.2.png
60fps-0.3.png
LG PF50, LG PF1500, RangeVision DIY: 2x DahengMer630, 2X12 and 2X16 mm 5Mp ZK lenses, RV turntable
Post Reply