How to remove single hot pixels in raw color image of QHY8

[oldwexi]

Usually i am exposing 2 to 3 and more nights on one object.
This leads to lights and darks with different hot pixels on different locations.
Before the life with PI i solved this issue with different SW which allows
the remove of isolated hot pixels on raws before calibration very efficient and easy.
So i could forget the darks anyhow as the QHY8 has very very low noise. Worked with
lights, flats and bias only.
After the jump into the PI-Pool and learning how to swim like a piece of iron
i wanted to get rid of the way to use for each image processing step another pool...
So far i did not find a function in PI which allows the remove of single hot pixels in raws.
Long opening,

The question is: Is there a simple way to remove
hot pixels in the raw (BW) color light frames with the help of PI?

Did a search in the forum but the search was not successful.

Thanks for your help
Gerald

[Jack Harvey]

Do not think you can do this before calibration step in PI.  Good darks should solve most of the problem with hot/cold pixels.  Other than that, I rely on the pixel reject algorithms in the Image Integration tool to take care of this issue.  This will require some experimentation to get the proper settings.  The use of these rejection algorithms assumes you dithered your sub exposures, which is what I believe most of us do during data collection???  Is there a reason you want to remove hot pixels before calibration?  I know certain software allows this and the workflow of some tutors emphasizes this, but I am not sure of the advantage of doing it at this step and not a bit later?  If you search they too often recommend dithering and allowing sigma reject to deal with the hot pixels.  HTH?

[oldwexi]

Thanks Jack for sharing your experience!
Will try to make better darks per night and calibrate the lights per night too. Because out of my experience
the location of hot Pixels differ to much from night to night...

I am using a STV for guiding. I did not find a dithering mode with the STV.

Gerald

[Silvercup]

Hi:

What about Cosmetic correction script by Nikolay?

Best. Silvercup

[oldwexi]

Hi Silvercup!
Thanks for your hint. Will try with the Cosmetic Correction to fight the raw hot pixels.
Gerald

[Jack Harvey]

Often dither you will find is a function of the imaging programs such as Maxim, CCDAP etc.  Dithering can be done by either moving the guide scope or the mount.  If you use Maxim for camera control then you can set it to dither automatically between images.  However, one can also dither manually between exposures (that is what we do when running the 1.23 meter at CAHA).  At the end of an exposure move the scope a couple of pixels (1-3 or so) in both the x and y axis.  ie image1>move x 2 and y 3, Image 2>move x-3 and y-1, etc.  Yep its a pain<G> and Cosmetic Correction is a great suggestion.

[jmtanous]

Hi,

I am also interested to hear on a possible solution for this problem.

I have a SS pro and I use nebulosity to remove the hot pixels of my lights before calibration.

The workflow that produces better result for me is:

+ Capture images with nebulodity (by the way nebulosity plays nicely with PHD, you can dither your lights automatically)
+ Remove hotpixels using a bad pixel map funciotn in nebulosity
+ Calibrate my lights using nebulosity
+ Debayer my calibrated lights using nebulosity
+ Align the lights using PI
+ Stack the lights using PI

I did a lot of exprimentation to try to calibrate my lights using PI, but the results were noisier no matter what.

Cheers,

Jose

[Jack Harvey]

I am not much help here as I have no experience with one shot cameras, debayering etc.  Sander?  Others?

[Harry page]

Hi

I use my sxvf m25 osc with pixinsight with great success , this is the same chip as the qhy8 and as said it is very noise free  so I do not use darks as all 

Because of a problem with Astro Art files I do not use Pi to calibrate my images at the moment , I await a new improved version  of PI

So I use AA4 to calibrate and debayer my images , and make no effort to use a bad pixel map or any other defect repair  ; I then bring the images into PI

Now as the main man says if you have dithered your images ( very very very recommended ) using Pi integration and a pixel rejection routine ( as this vid http://www.harrysastroshed.co.uk/Stack.html ) I find it will remove 99% of my outliers.

As I am a simple man ( Stop nodding you head ) this is a simple and effective way of doing things , why complicate it with all the other stuff , let PI take the strain

Harry

[mmirot]

I suggest you still subtract a bias.

Max

[Harry page]

Hi

Of course if you are using flats a bias would need to be used on the lights, but if you are not using flats I do not find you need to subtract a bias from the lights , this only works with my sxvf m25 
I would always subtract a bias from my sxvf H16 as it has some chip properties that can be removed this way ( i.e. Banding )

Harry

[oldwexi]

Hi Harry!
Like Jose, i wanted to remove the hot pixels on the RAW images before calibration because i more
or less use the same procedure as Jose for calibration and image processing.
Certain SW allows to remove the tiny single RAW hot pixels without destroying any real data.
Doing the hot pixel remove after debayering i got the experience that the tiny RAW Hot pixel
after debayering became red, green od blue dots which are not so easy to remove than the
raw single hot pixels.
Yes,  the QHY8 is very noise free. So i only do bias and flats, no darks.
And as mentioned instead of the darks i wanted to use the raw hotpixel removal.
Darks and/or  bias do not work for complete hot pixel removal. Because the darks
change over the days and the warm/hot pixels move to different locations from one day to the other.

So a PI SW solution for removal of single tiny RAW hot pixels before calibration would be nice.

Dithering with an SBI STV guider is not working for me as the STV is a standalone device and has no interaction
with dithering/guiding software.

Gerald

[Harry page]

Hi

I am afraid I do not know of a way of doing this in pixinsight , Mr Jedi wil have to put up here

But please do look at dithering in future 


Harry

[Nocturnal]

Hi,

please please please don't confuse noise with dark signal. The QHY8/M25C have virtually no dark signal but that doesn't mean they are low noise! I don't mean to be a pain but terminology matters when you're discussing such matters across languages and time zones

So, that said. My recommended way of stacking QHY8 images is as follows:

- remove hot pixels and bad columns if any
- stack lights with flats and bias (I use DeepSkyStacker)
- process the resulting 32b TIFF in PixInsight

For hot/cold pixel and bad column removal I wrote a free utility you're welcome to try.

http://www.tungstentech.com/Software/FixFITS/tabid/78/Default.aspx

Eventually I'll port this to PCL but it's a bit like using an alephant gun to kill an ant. I'll get to it but it's low priority as FixFITS doesn't have many users and it's already available for free.

[jmtanous]

Sander,

Your FixFITS utility does not run on my system (virtualized XP on OSX). No big deal cause I use bad pixel maps for hot pixel removal, but it will be great if you port your tool to PI, that way I can run it natively on OSX.

Cheers,

Jose