OSC red and blue hot pixels in integration result

Hot pixels normally get identified when using darks but you already do that. Can you zoom in to a region with those hot pixels? All I see in your image are very red or blue stars, which is odd, but those aren’t hot pixels because they cover more than one adjacent pixel.

Wouter

These still look like hot pixels though, if they are stars that would be very weird. With darks, bias and flats you can make a nice bad pixel map btw, just select it when you have loaded in all the calibration files and use it during the process (also for later sessions). Could you upload a 1:1 picture here (jpeg form is fine as it has to fit 30 MB) so we can zoom in nicely?

Vincent is right. These are not stars. They are hot pixels. 

Discussed this with QHY technical support, they are suggesting that the offset value is too small to create good calibration frames (dark and bias frames). QHY suggests that the offset value should be higher to prevent calibration frame from clipping. That's next I am going to try.

Just trying to see if there is a way to save the existing data. Especially if I want to do Ha-OIII/Ha extraction and Ha-OIII/OIII extraction, this will be more important. 

Vincent, are you asking for 1:1 calibration master frames or the integration result? I can upload them when I get home. 

Thanks!

Did you try do perform some, maybe a bit aggressive, sigma clipping? Like LN MAD and then kappa 2.0 and 2 iterations. It may be less nice for some details, but if that helps to get rid of them, it might save the frames. Did you dither when capturing? That will help as well.

Just a 100% jpeg of the result as you posted above would be nice.

Thanks Vincent!

No I didn't try with Sigma Clipping. Actually that's what I want to learn from here. I will definitely give it a test (Sigma Clipping / LN MAD / Kappa 2.0 w/ 2 iterations). Any more ideas that I could try?

As of dither, yes and no. I did dither every single frame. But I didn't adjust the dither settings from the mono setup so it won't work. The OSC scale on my system is 2.18 and the main camera scale is 1.708. I only did the small dither (0.5 pixel) and dither scale of 1.0. It's not even close to remove the hot pixels. That's another thing I want to see how it could help - pushing up Offset value and increase dither size. Only concern I have is how much time it will cost the mount to settle with large dithers. 

I will post the 100% JPEG when I get home. The picture I posted is actually 1:1 crop from the integration result. 

I'd try the clipping first, might already help quite a bit.

Dithering will help a lot! Especially if you dither randomly and maybe at normal to somewhat higher setting. Yes, it will need to settle, but my experience is it usually takes about 5-10 seconds for a high dithering. I only do that every 2-3 frames and since I know the benefit, I'm happy to take a bit of time for it.

Vincent, I am happy to report that following your suggestion the red and blue hot pixels are now successfully removed. Here is 100% comparison. 

Thanks! 

Want to ask a quick one that is totally not related to hot pixel. How do I change the "greenish" O3 color to nice blue color in APP? 🙂

Great! So now you can try to dial that setting back to maybe kappa 2.5 or 3.0 with 1 or 2 iterations, so it will do less "damage" if at all seen.

The green you can dial back with the "HSL selective color" tool, I would try to change "Cyan", drag the G<->MA (green-magenta) to magenta a bit, B<->YE (Blue-Yellow) to blue a bit. Maybe the same on the "Magenta"...

I know this is an old thread, but I have bumped into this same EXACT problem with my ZWO 533 camera. Im so glad I found this post.

Please tell me the exact steps you took to eliminate the RGB pixel problem. It has wasted hours of imaging and image processing. After extensive trouble shooting (with APP and PI) I had almost given up on the 533.

Posted by: @vincent-mod

Did you try do perform some, maybe a bit aggressive, sigma clipping? Like LN MAD and then kappa 2.0 and 2 iterations. It may be less nice for some details, but if that helps to get rid of them, it might save the frames. Did you dither when capturing? That will help as well.

Just a 100% jpeg of the result as you posted above would be nice.

These suggested options come up in several places. Please specify which exact menus you are referring to. Thanks!

Glad you found some info here. So what I would suggest is to check your darks first, check if the offset is correct for this camera as the manufacturer also mentions it should be set higher (see message above). Then redo the darks and Bad Pixel Map and check if that makes it better.

If not, you can set the rejection parameters a bit higher; go to tab 6 "Integrate" and switch the integrate option from "automatic" to "average" (if you have more than 20 frames), then underneath that the filter option becomes available, set that from "no rejection" to "adaptive rejection". Then the Kappa high, set that from "3.0" to "2.5".

Continue with integrating and check to see if that makes a difference already. If not, you can lower the kappa high a bit further to "2.0" for instance. Ideally though, you would want this to be done by the darks already and by having a good dithering.

Thanks so much for the quick response!

According to the FITS, the camera is set (automatically by ASIair) to 70 which is the manufacturers suggested setting.

I just did the flat / darks / dark flats several weeks ago , but I will do them again in hopes it will help.

If that doesnt help I will try your very clear rejection settings. 

I sure hope this problem can be resolved.

Thanks again.

It will be resolved, it's not a huge issue for the software. But it's always best if you can get the data as nice as possible beforehand. I'd suggest to also make a long dark (like 10 min exposure) to maximize the hot pixels for the bad pixel map. This bad pixel map you can then use for a long time (at least a year I'd say). Still make the darks for your exposure of course, but the bad pixel map can always be added to any dataset.

@vincent-mod The darks I made a few weeks ago are 10 min. which is the same length as the Lights. And I did make a Bad Pixel Map from them.

What I cant figure out is, since I just made all the calibration frames(Flats / Dark Flats / Darks) for the 533 a few weeks ago (darks made specifically for the 10 min. Lights) why would I be having these hot pixel problems? It doesnt make sense to me.

Well, regardless, Im making some more 600 Darks since after all my research and troubleshooting, that is the most likely reason things are not working as they should. Im chiiln' my 533 in the refrigerator in the Vegas daytime. Its cold (easy on the camera's cooler) and its dark. It just might be a new ritual for me when needed.

I must say I have no experience with the 533, so there may be something very specific as the root of the issue. Do you dither though? That will also help a lot, dithering with 4-5 or even more pixels.

Regarding putting it in the fridge.. be careful with that as I can imagine moisture getting into the camera that way.

The interesting thing is Im having exactly the same experience with my 533 as the OP had with his QHY 168. Maybe they use the same sensor?

Living in the desert (Vegas), Ive never had a problem with refrigerator condensation. Lets hope it stays that way.