Red blue green pixels at centre of any image

Njp

Hi there
i haven’t had a chance to image for a long time here in the UK due to bad weather but tonight I went out and used the exact same setup as all times before but got bad results. The centre part of any target, for example M31 had lots of red, blue and green pixels. The effect was less at the edges and decreased the further out from the centre. I used an ASI294MC Pro camera with a William Optics GT81 scope. Cooling was on to -10 and guiding was very good, better with the latest firmware I think. Does anyone have any ideas as to what the problem would be. The optics looked clean and camera sensor looked ok to me. Image of example attached. Thanks.

w7ay

Njp The effect was less at the edges and decreased the further out from the centre.

If the center is worse than the corners, the first thing I would suspect are tiny ice crystals on the sensor. Check if they are there if you hadn't turned on the camera's cooler.

If they are not there when the camera is not cooled, then it may mean that it is time to install new desiccant tablets.

By the way, if it is ice, they should not be Red Green and Blue pixels, but Cyan, Magenta and Yellow pixels (with different degrees of saturation -- e.g., yellow could appear brown, etc). E.g., the crystals are preventing red wavelength to pass through the Bayer filter, an thus creating a magenta superpixel.

Unfortunately, it is hard to tell from a JPEG image which compresses chroma. You will need to inspect the original raw image.

By the way, with a few dozen subframes, dithering should reduce the problem. But it is better to fix the root cause, instead of trying to mask the problem.

Chen

Tech@ZWO

Njp The picture is compressed, can't zoom in.
Looks like the dew, can caused by the high humidity inside the camera.
You can cool the camera slowly to try.
Cool to 5℃ first and keep for 10 minutes, then cool to the target temperature you want.

Njp

Hi Chen, thanks for your help. Did you zoom into my image to see the coloured pixels? My cooler was on and set to -10. Temp was -10 when image was taken. I abandoned imaging as it did not get better as time passed. I followed same setup as always and have never had this problem previously. I couldn’t see any condensation or ice on the sensor window after I finished. Outside temp was 12 degrees C.

enaiman

Njp I suspect the dark/bias frames you have used for stacking. I think those pixels are everywhere but more noticeable in the center because your image is the brightest there.

w7ay

Njp Did you zoom into my image to see the coloured pixels?

Yeah, I did and could not actually measure the color of each subpixel since it has gone though JPEG compresssion.

As I said, you need to pixel peep at pre-compressed image. If you see for example lots of Magenta instead of lots of Red, then odds are they are causes by tiny ice crystals (remember that each pixel is only a couple of micrometers -- you won't see them).

Sometimes, a good sensor cleaning could also help keep the ice from forming.

The reason I suspect ice is that, just like the dewing problem, they tend to start at the center of a piece of glass.

But you can easily proof or disprove that it is caused by ice crystals by not turning on the cooler and take some images. Then start the cooler and take a few more frames.

You can actually do it with illumination from a flat panel on a cloudy night. But I would try to give it the same environmental conditions, and thus perhaps not indoors (although you can try that too, to see if the chromatic spots also appear).

Njp

enaiman Hi enaiman,
Thanks. These shots were single frame captures viewed on iPad connected to my ASIAIR, so no dark or bias frames involved. I agree with what you noticed, the artefacts are more visible at centre due to lighter image there. I have never seen this before when viewing single captures. Only difference from last imaging is I have upgraded to V2 app.

Njp

w7ay
Thanks Chen. I will try again when weather ok as my setup is on a steel pier in the garden with power etc in place so difficult to test without being outside etc.

w7ay

Njp I will try again when weather ok as my setup is on a steel pier in the garden with power etc in place so difficult to test without being outside etc.

Sweet.

Unless it is raining, perhaps you can try it even when it is cloudy, by using a flat panel (even an iPad that is displaying a white image as the flat panel :-).

I don't have a permanent pier myself, but do have a permanent tri-pier (a WO Mortar) outdoors, with power and Ethernet cables led to a ABS box at the base of the tri-pier. But I have also replicated everything (same 12V hub, same USB hub, same mount, etc) on a carbon fibre tripod indoors for doing my indoor tests (with artificial stars). The indoor setup doubles as my portable mount, and it also serves as "storage space" for an OTA :-).

Chen

Njp

w7ay b
Your setup sounds good. I will try to image again soon in between bad weather !

I am a little puzzled about why the ice was forming as outside temp was around 12 C and my cooling was set to -10 C as I normally do. Nothing was different from the last time I imaged apart from being on V2 of ASIAIR software. Can you explain why you think it is Ice ? BTW as per previous post by enaiman I think artefacts are not just in the centre area, just more obvious there as it’s a lighter part of the image.

w7ay

Njp Can you explain why you think it is Ice ?

Because (1) they are really tiny (typical dust particles are larger -- the air quality sensors are based on 2.5µm and 10 µm particles, for example -- about the size of a pixel). (2) they are more concentrated near the center.

Anyway, easy hypothesis to test by taking images with the cooler off.

If it is not ice, the next candidate is some sensor flaw. Hope not.

Chen

Njp

Thanks for your help Chen. I managed to stack using live mode with my ASIAIR Pro for 30 mins without cooling initially and all was ok with first 4 subs so I switched cooling on and set to just 0 degrees instead of -10 C and did 6 more. The final stacked image is attached. It was much better with no artefacts like I had before. What was impressive was a satellite trail on sub 4 that gradually got removed as the rest of the subs got stacked as the live mode progressed. I guess I didn’t give the equipment enough time to acclimatise last night as tonight I set it up and waited an hour before starting. Thanks. BTW. Very glad it didn’t turn out to be a sensor !

w7ay

Njp What was impressive was a satellite trail on sub 4 that gradually got removed as the rest of the subs got stacked as the live mode progressed.

Good post processors have something called sigma clipping, where you can remove a satellite trail with as few as 3 subframes.

Imagine a pixel at some (x,y) location. When you stack three subframes, you typically would take the mean (average) of the (x,y) pixel of the three frames. However, instead of blindly averaging, you can also compute the standard deviation (sigma, which is the square root of the variance) of the three values. Any pixel that is outside a fixed standard deviation from the mean can simply be discarded. So, instead of taking the averge from three frames, you simply discard the pixel value that has the satellite, and only average from two frames. The noise along the satellite path would be higher, but you will never notice it, especially when you use a larger number of subframes than 3.

By the way, I took your image into macOS Preview.app, and appled an automatic white point correction using a pixel near the center of M31 core (avoiding the saturated pixels, which are already white), and get this:

If you like pixel peeping a large image (40 MB), here is one I took of M31 with an FSQ-85 and stacked with AstroPixelProcessor (camera angle 90 degrees from yours). Exposure was kept short so that most of the stars maintained their colors.

http://www.w7ay.net/site/Images/FSQ85/M31.png

Chen

Njp

Thanks for the explanation of how it works. I’m very new to the hobby, had equipment for 2 years but here in the UK the weather is generally either cloudy, raining or windy, or combinations of these so difficult to get out a lot. Thanks for the tip about star colours being preserved with shorter exposures, I didn’t realise that, so will try the next time I’m out. I only did a small live stack of 30 minutes as clearoutside app said there would be some clouds but it turned out it was clear after I had packed up my gear! Next time I will try shorter exposures but a lot more of them. Thanks.