Machine vision cameras

DMS General
#1

Anyone here familiar with machine vision cameras?

I’m planning to upgrade the camera behind my x-ray machines phosphor plate. The machine operates by sending a series of x-ray pulses to illuminate the phosphor plate. Depending on operation choice, the camera can do a long exposure of multiple x-ray pulses or several separate exposures and be summed in software. An upgraded camera can allow for brief moments of burst video with rapid pulses synced with the refresh rate.

Here’s some samples of its current results (yes, I know I could screenshot but there is some wonky stuff going on that I have to deal with first)

PXL_20210910_214309235
PXL_20210910_2143092351512×2016 1.3 MB

PXL_20210920_063219319
PXL_20210920_0632193192016×1512 1.28 MB

Current camera has a f0.8 aperture. I’ll post more details on internals when I get it apart next

1 Like
#2

It could be the camera is fungible or it could be a specific model for reasons that aren’t intuitive. I wouldn’t change it if you can avoid it. Or try to find the same model or at least the same specifications.

FLIR bought Point Gray a few years back. I wouldn’t be surprised if the camera in the machine right now is from Point Gray.

https://www.flir.com/browse/industrial/machine-vision-cameras/

#3

We use Intel Realsense cameras for CV at work.

Lol, is that a suppressor in the bottom pic?

1 Like
#4

Yep. For a .22

#5

The hardware hacking community has already been replacing the cameras, but done a poor job of documenting the changes and how they selected them. There isn’t a reason to use the original in-unit.

For some background this machine was built in 2004, and even then the camera wasn’t “what it could have been”

They camera and it’s control assembly remove from the unit easily and can always be out back in.

#6

The most detail I’ve gotten so far on any upgraded units is from the RTF Technologies guy, however there’s only the video explaining the upgrade and no samples: SAIC RTR-4 x-ray Integrated Imager camera upgrade - YouTube

However they did call out a couple of specific camera configurations:

Basler ace acA1920-40gm

  • 5.86 micron square pixels for 60% sensitivity in green spectrum
  • Kowa LM8hc 8mm f1.4 lens

Basler ace acA3800-10gm

  • 1.67 micron square pixel w/ 40% efficiency in the green spectrum
  • C125-0618-5m lens 6mm @ f1.8 (not as good light gathering but a nice match for the system overall)

I’m hoping to be able to scrunge around and play with things a bit before splurging on a specific arrangement. Machine vision cameras are something I don’t have as much experience with (at least that isn’t dated longer than 10+ years)

#7

I’ll be at DMS today and dissecting the camera head a bit for some measurements and testing (and testing a new dosimeter over in science).

I won’t be powering on the generator head because I think that’s not appropriate to do there, but if anyone wants to check the machine out can show you how it all works.

1 Like
#8

Things came up and I have to try again; always something that needs handling at DMS.

I’ll be back tomorrow with it and try to crack it open after my lift class.