Line pairs per mm, resolution, and a Raspberry Pi camera module with a Canon lens?

I've been experimenting with a bit of astrophotography (5D4 on a tracking mount, with a 70-200 f/2.8 IS II). Obviously this doesn't give much "reach" for galaxies, so I've knocked together a mount that allows me to put a Raspberry Pi V2 camera module (with its lens removed) on the back of the 70-200 lens. My basic question is: is there any point in continuing with this?

The Pi camera module uses an 8MP Sony IMX219 sensor, with a diagonal size of 4.6mm and a pixel pitch of 1.12um.

The sensor on the 5D4 has a diagonal of around 43mm, and a pixel pitch of 5.36um.

I make the crop factor 43/4.6=9.3x. Therefore I should get a framing roughly equivalent to 200x9.3=1,860mm when I use the 200mm lens on the Pi camera.

However, will I be getting enough detail from the lens to make this worthwhile? Based on my (weak) understanding of MTF data and lines per mm (lpmm) values, the gist I get is that the 70-200 is roughly 40 lpmm at 200mm and f/2.8. If this (https://www.optowiki.i​nfo …millimeter-to-pixel-size/) is accurate, then that 40 lpmm is probably not out resolving even the 5D4 sensor (40 lpmm "supporting" a 6.3um pixel size, with the 5D4 already being finer than this at 5.36um). Based on that, there's no way the 70-200 is going to be resolving enough detail to make use of the Pi camera's 1.12um pixel pitch.

However, I understand that these things aren't black and white, and that having the sensor out resolve the lens is not necessarily a bad thing.

I could just crop the 5D4 captures to match the sensor area of the Pi camera, resulting in approx 720x480 (0.35MP) images. What isn't clear to me is if the equivalent capture from the Pi camera (at 8MP) will actually be better - especially once a stack of images are processed to produce on final output?

With a 70-200mm lens the best and probably the only galaxie you can image is M31 which 200-300mm is about right size for a lens . Other than that any other galaxie i doubt seriously you can get anything especially detail . Maybe the Markarian's Chain of galaxies which all will be very small and fussy .

Celestron wrote in post #18974127
With a 70-200mm lens the best and probably the only galaxie you can image is M31 which 200-300mm is about right size for a lens . Other than that any other galaxie i doubt seriously you can get anything especially detail . Maybe the Markarian's Chain of galaxies which all will be very small and fussy .

Absolutely - but in theory, a 200mm lens that could resolve a (probably impractically high) number of line pairs per mm could be used to project an image onto a tiny but high resolution sensor; thus giving the field of view of a much longer lens. The question is more if the 70-200II projecting onto the tiny area of a IMX219 sensor would be worth it. Either I'll just be highly sampling "mush" (meaning I have 8MP of image that contains no more detail than simply cropping the full frame image to the same field of view) or it will contain more useful detail.

I'm thinking (possibly incorrectly) that multiple exposures from the IMX219, once stacked, may result in a better level of detail and noise vs a heavily cropped 5D4 image - but I could be wrong (hence the question).

sploo wrote in post #18974352
Absolutely - but in theory, a 200mm lens that could resolve a (probably impractically high) number of line pairs per mm could be used to project an image onto a tiny but high resolution sensor; thus giving the field of view of a much longer lens. The question is more if the 70-200II projecting onto the tiny area of a IMX219 sensor would be worth it. Either I'll just be highly sampling "mush" (meaning I have 8MP of image that contains no more detail than simply cropping the full frame image to the same field of view) or it will contain more useful detail.

I'm thinking (possibly incorrectly) that multiple exposures from the IMX219, once stacked, may result in a better level of detail and noise vs a heavily cropped 5D4 image - but I could be wrong (hence the question).

Good luck , hope your experiment works well for you . Have a great day !

I am no expert, but I looked at MTF charts for this lens and not sure about the conclusion "that the 70-200 is roughly 40 lpmm at 200mm and f/2.8". On the OLAF chart, 50 lpmm is still .6 at 2mm from the center, which would suggest that the lens will do much better than 40lpmm. So the Pi sensor might be worth a try. I am curious about the adapter/mount that you plan to use.
Best,
Steve

stevieray wrote in post #18979119
I am curious about the adapter/mount that you plan to use.

Here's the current (very rough) prototype. It's just a Pi camera board screwed into a small piece of wood, held onto a Canon lens cap using a pair of small bolts. The black line on the wood indicates the plane of the Pi camera sensor; it's important to get this exactly 44mm from the lens flange (to match the distance between a DSLR sensor and an EF (or EF-S) lens).

The small white plastic tube isn't structural - it's just there to block stray light getting onto the sensor.

Hopefully I can get away with breaking the rule about uploading photos in consecutive posts as they're quite small...

With the body cap removed (leaving just the lens cap) you can see the hole cut into the lens cap, and the Pi camera sensor. The Pi lens has obviously been removed - using the little white focusing tool they provide with the camera.

The head of the bolts are just shallow enough to not foul the lens flange.

Take care if you drill out a hole in one of these lens caps - the plastic can be quite brittle and "grabby"; which isn't fun if you're using a drill press.