I did some quick research on the aggregate positional accuracy of laser printers. I see your point but I'm still not sure just how much error that is inducing into our scans. That said (and because I have nothing better to do) I'm going to try to make the most accurate panels possible. With the resources I have at hand I guess the best way will to use a VMC to mark/mill the panels. I happen to have some 1/4" ground fixture plate left over from a job which should provide a very flat starting point. On the large VMC I should be able to make 200mm panels with .0002" (.0051mm) positional accuracy.bigbomber wrote: ↑11 Jan 2018, 23:33
- if vertexes of projected squares aren't lied exactly on the center of circles but you see jagged edges, this is the result of incompatible projector (or can be caused if cameras can't see calib panel correctly (Dark Image, Not enough contrast & Brightness, unfocused lens & etc)).
- if vertexes of projected squares are lied exactly on center of circles but you see jagged edges, this is the result of not-accurately-made calib panel.
In brief, software try to project squares vertexes to centers of circles, even if they are not in their correct position. but if software can not calibrate projector correctly, efforts will be failed and vertexes deviated from centers.
this is my idea. what do you think.
I'm going to create a test on my idea. can you guess the method? it is simple.
Meanwhile, It is very nice to see the forum come to live again. thanks to all friends.
Projectors... and Projector Mods.
Re: Projectors... and Projector Mods.
µ
Re: Projectors... and Projector Mods.
I understand the meaning of diamond pixel type. But my projector NEC is not expensive and has squre pixel (edge of pixel lied in H&V direction).
Re: Projectors... and Projector Mods.
I'm wondering if you rotate your calib panel enough to do scan as second position. It must be rotated at least 15 deg to show any deformation.erapip wrote: ↑11 Jan 2018, 20:12 @bigbomber
I try to optimize and test with single camera with a positive calibration board. Scale is 120mm and Total angle is 24deg. See attache results. Its better than with dual camera. What do you think?
I have to ask about the FlexScan that you use before, was it better than David? Did you compare some scans? What about your experience with uEye camera?
About flex: I think single scan quality of HP is better. HP will project more patters to cspture scan. Also HP is more flexible and userfriend.
Anout diy equipments, I belive HP calib panel is easier to make according to its circular patterns (than squre patterns of Flex. If you look at high-end industrual scanners, all of them use circular patterns too).
About cameras, newer versions of Flex only accept expensive PointGray cameras.
HP loads alot of RAM during scanning, Flex is bettrer at this point. Flex also accept markers for alignment. It helps for automation scanning process.
Scanning process is faster than HP (note flex projects less number of patters). Flex alignment process is a little faster.
And I can cature more details using HP sith equal equipments ans setup.
Re: Projectors... and Projector Mods.
I think so: see picture (the same I posted about the 220mm pattern): please look at it carefully (you probably need to zoom in).
- Attachments
-
- Diamond.png (35.39 KiB) Viewed 10273 times
LG PF50, LG PF1500, RangeVision DIY: 2x DahengMer630, 2X12 and 2X16 mm 5Mp ZK lenses, RV turntable
Re: Projectors... and Projector Mods.
Why do you think that a jagged line on a diagonal edge is cause by a projector problem and not just rasterizing due to projector resolution?
And What do you do to a projector to solve this?
µ
Re: Projectors... and Projector Mods.
I'm agree there is jagged as a result of rasterizing. But in a good calibration it is very small and hard to see by eye.
-
- Posts: 11
- Joined: 07 Mar 2017, 07:47
Re: Projectors... and Projector Mods.
I should mention that this can be achieved only in case of projector is perpendicular to the calibration board. In other cases there shall be jagged edges because of parallax effect.
Re: Projectors... and Projector Mods.
Friends. I'm interested to know if patterns count for all of us is the same in "Quality Mode" and "Speed Mode".
Can you Check Please?
mine are 60 & 40'
Can you Check Please?
mine are 60 & 40'
Re: Projectors... and Projector Mods.
I'm scanning on default which is 33 patterns. I have it set for Horizontal only.
µ
Re: Projectors... and Projector Mods.
Yesterday I got some time in the shop so I had a very close look at my calibration panels. I was surprised to find a number of errors and problems. Without remaking the panels, which is what I should do, I adjusted and fixed as best I could. The results were far better scans when calibrated with these panels (225mm). When I get some time I'm going to have to look more closely at all the other I have. Live and learn.
@bigbomber I see what you are saying about the jagged edges. the better the calibration the less obvious and present these edges seem to be. I still don't exactly under stand the mechanism behind this, or what in the projector is causing this.
I tried your panel experiment with after I "fixed" my 225mm panels and got better results but still not great. However when I scanned a small casting the scans were pretty good.
Another thing that seems to help is using an inverted corner.
It is interesting that one side seems flatter than the other. A Note. I'm scanning with my laptop here and not the computer I usually use. I haven't gone through the video settings on the laptop to match the tweaked setting I have on the other machine so that may be part of it.
Here are the actual scans I did.
@bigbomber I see what you are saying about the jagged edges. the better the calibration the less obvious and present these edges seem to be. I still don't exactly under stand the mechanism behind this, or what in the projector is causing this.
I tried your panel experiment with after I "fixed" my 225mm panels and got better results but still not great. However when I scanned a small casting the scans were pretty good.
Another thing that seems to help is using an inverted corner.
It is interesting that one side seems flatter than the other. A Note. I'm scanning with my laptop here and not the computer I usually use. I haven't gone through the video settings on the laptop to match the tweaked setting I have on the other machine so that may be part of it.
Here are the actual scans I did.
µ