Hot pixel removal without darks

[fmazzanti]

Hi folks,
Just a quick question: I use a. Trius 694 camera that has a Sony chip that has so low dark currents that no dark frames are needed (or so they say). So I don't take darks, only lights and bias. But in 1x1 bin mode I get some hot pixels that I'd like to remove in the final image... Can somebody tell me of the best practice to remove them, considering the lack of dark frames?
Thanks,
Ferran.

[pfile]

you can make a master dark out of a few dark subs, then use it as input for the CosmeticCorrection process. you can use the sliders to essentially isolate the hot pixels and then apply the process to your images. you need to load some images into the file dialog of the process, plus the master dark. then you can open one of your subs, define a preview on it (just so you can see 1:1) and then turn on the realtime viewer and select the preview in the RT view. as you change the sliders you'll see the hot pixels disappear... try to get the slider such that it's only removing obvious hot pixels...

when you execute the process it will apply the correction to each of the files loaded, and write them out to disk with a _cc postfix.

rob

[Geoff]

Harry describes this process here http://www.harrysastroshed.com/pixinsight/pixinsight%20video%20html/pixinsighthomenewbie.html.  Look under "Preprocessing".  You can also use the defect map described here http://pixinsight.com/doc/tools/DefectMap/DefectMap.html
Geoff

[sreilly]

I guess the question becomes what should be done when. Usually I calibrate the images using master dark and flat master frames. Now I could see using Cosmetic Correction although I still don't have a feel for using it but then I also wonder if combining the images doesn't get rid of a good amount of these. But I guess you're better off getting rid of as many as you can before combining.

Next thing I need to do is see how I can use the defect maps I made in MaxIm to get rid of the partial column defects. I thought I could use the same X/Y coordinates but I think the image is orientated differently in PI as they don't seem to work. any suggestions?

Steve

[fmazzanti]

I must read all that but... Does 't all these methods require ysing a master dark, or a few dark subs at least? I was looking for some sort of solutions that do not require darks *at all*. I was able to do that in Astroart very easily, and actually I can imagine a simple interpolarion of the surrounding pixels could do the trick. Is there a way to do that, really?
Thanks again,
Ferran

[Andres.Pozo]

CosmeticCorrection doesn't need dark frames. It can use them but it can work very well without them using the "Auto detect" mode.

[cfranks]

Steve,

You can use CosmeticCorrection process to get rid of those partial Hot Columns.  With 'Use Defect List' checked and 'Col' checked,  Mouse-over the Master Dark or image hot column, click will select it then Add Defect puts it in the Defect List. 

Charles

[sreilly]

Thanks Charles,

I must have missed that before.

-Steve

[pfile]

I must read all that but... Does 't all these methods require ysing a master dark, or a few dark subs at least? I was looking for some sort of solutions that do not require darks *at all*. I was able to do that in Astroart very easily, and actually I can imagine a simple interpolarion of the surrounding pixels could do the trick. Is there a way to do that, really?
Thanks again,
Ferran

i guess if you care so little about the quality of your data that you are not willing to make a few darks, then why bother with any of this?

if there are enough subs and they are dithered the hot pixels will be rejected during integration anyway, as long as you have chosen a rejection algorithm.

rob

[sreilly]

I agree. Taking darks is simple and can be done when it's not so good for imaging so why not and at least see what difference it makes? Forget what other software allowed you to do. Maybe with the darks you'll be able to stretch the data more and dig out dimmer details.

-Steve

[fmazzanti]

i guess if you care so little about the quality of your data that you are not willing to make a few darks, then why bother with any of this?

Uhmm... nice way of putting things.
My camera (starlight xpress Trius 694) has a Sony chip that has so little dark currents that even the manufacterers say that darks are not required, mostly considering that darks, as any other image, will add noise to the result anyway. And darks depend on temperature so it's not as  simple as taking a few darks and you're done, but building a library... Or so they say as I have never done that myself, so it is not my experience.
I love avoiding darks considering I have to pack and unpack a lot of stuff everybtime I take images, which is not that often... And bottom line is that, as I said, Astroart can do a very nice job at killing these pixels with zero effort, so I was hoping for something similar in PI. But I see it is not that simple in PI ...

[topboxman]

CosmeticCorrection doesn't need dark frames. It can use them but it can work very well without them using the "Auto detect" mode.

+1 what Andres said. I have used Cosmetic Correction without any dark frames and works very well. My camera is also ICX-694 based (QSI 660wsg).

Cosmetic Correction is very simple but can be confusing for first time user.

Peter

[pfile]

i guess if you care so little about the quality of your data that you are not willing to make a few darks, then why bother with any of this?

Uhmm... nice way of putting things.

look, my point is that you only need a handful short darks integrated together to get a good map in CosmeticCorrection. as others have pointed out you don't actually need the dark, but i figure why not use the dark to guide the CC process? the dark contains the truth. you can guess which pixels are outliers (which is all AstroArt and even CC is doing without the dark) or you can just go right to the source.

taking one or two darks every once in a while is a far cry from making a 50-100 frame dark master or taking a billion darks every night. so i don't really see why it's a huge burden.

by the way these sensors are stable enough that you can probably use the same dark for months and it will be perfectly fine for the purposes of identifying hot pixels.

rob

[fmazzanti]

CosmeticCorrection doesn't need dark frames. It can use them but it can work very well without them using the "Auto detect" mode.

+1 what Andres said. I have used Cosmetic Correction without any dark frames and works very well. My camera is also ICX-694 based (QSI 660wsg).

Cosmetic Correction is very simple but can be confusing for first time user.

Peter

Aha, thanks... I can try that and see what can it do for me. Have you tried also with the dark frames, to check what real difference it does with that chip?

Thanks a lot,

Ferran.

[fmazzanti]

I agree. Taking darks is simple and can be done when it's not so good for imaging so why not and at least see what difference it makes? Forget what other software allowed you to do. Maybe with the darks you'll be able to stretch the data more and dig out dimmer details.

-Steve

Sometimes is not that easy... In my case I have to pack/unpack most of the equipment every time I take images, so doing that when t's not that good for imaging still takes a bit of time. I wish I had a permanent setup to do all these things and others) but unfortunately that's not the case

Still if it is just a few darks that one needs I may give it a go next time and check, because your point about being able to stretch more the data sounds interesting.
How many of these do you need? I usually take 10min exposures or so, so already 3 makes it half an hour... what would be the right amount of darks to try the cosmetic correction thing?

Best regards,

Ferran.