›

Black pixels added to image after Calibration with WBPP

Still_Cloudy
in PixInsight
Hello,

I'm new to Pixinsight and am trying to master it, but I haven't been able to determine what I'm doing wrong in this instance.
I've looked at the images produced in Pixinsight after running Calibration, Cosmetic Correction and Registration in the WBPP script, and found that there are black/dark pixels throughout the image that weren't there originally.  I've now rerun the script, only doing Calibration, and found that this is where the black pixels are being introduced.
I've tried various settings:
- In Calibration Settings, I've had both Dark and Flat set to Auto, and I've run it with and without Optimize Master Dark checked.
- For Output Pedestal Settings, I've run the script with mode being Auto; and other times, I've set mode to Literal Value to 100 and I've also tried it at 800.
In all cases, black pixels are introduced to the image at this stage.  The curious issue is that it seems that one black pixel is appearing next to each hot pixel.
All subs are correct in terms of exposure length, binning, temperature, mode, gain, etc. and the script runs perfectly every time.  I'm using a QHY600M camera.

I would appreciate any insight as to what I may doing wrong, and how I can correct this.  Thank you.

Bob H.

Comments

  • Still_Cloudy
    I've just now noticed this statement in the Pixinsight Process Console:

    * Performing image calibration.

    ** Warning: negative or insignificant pixel sample values detected after calibration.

    Applying automatic pedestal: 1000 DN (huge!)


    Is there a way for me to correct this?


    Bob H.



  • BlockHeadBlockHead
    Given that you are using a pedestal (good job!)- we can probably eliminate this as an issue.
    The DARKS have to be the issue. The darks are subtracted from the lights ..they are the only things that can make the dark spots.

    Your observation that the hot pixel do not align is important...but they are not misaligned in the same way. I do not know to achieve this. You should be able to blink a raw light and a raw dark...and the hot pixels should be in exactly the same position. It is as if the dark has been manipulated in some way...but I have no idea how.

    I guess I would need to see a raw dark and a raw light frame and the masterdark.

    If I had to guess...I wonder if you some how mixed dark frames when creating the masterdark. I have no idea. Please make the image available via a cloud server. https://www.adamblockstudios.com/articles/how-share-files

    There is something about the dark frame(s) that isn't right.

    -the Blockhead
  • Still_Cloudy
    Hi Blockhead,

    I will experiment tonight with the data, and will get the frames to you tomorrow.  You do have me questioning whether I used the correct darks.  Thank you for taking the time to help me with this; I do appreciate it.

    Bob H.
  • Still_Cloudy
    Hi,

    Attached is a link to images that I'm using.
    It contains:

    - Original light frame
    - Calibrated/Cosmetic Correction/Registered light frame
    - Original dark frame
    - Master dark frame (3 versions.  I don't know how these were created)

    Hopefully the answer to the black pixels is in here!

    Thanks again,

    Bob H.
  • BlockHeadBlockHead
    edited July 2024
    Something is clearly different. I cannot say why though. 
    It is clear the light frame and dark frame are not contemporaneous. You might complain, "but Adam" you might say  "they are only 2-3 months different in time?" Yes, darks "age" in diverging from characterizing the hot pixel population...sometimes quickly. So you should *always* take darks as closely in time as possible to the lights when doing this kind of diagnostic analysis. 

    Something changed in time....

    It shouldn't matter... but in the past it can! This is VERY difficult to capture, but if you change software that you used to acquire the images or EVEN the same software but different versions this can impact the camera driver and affect the results!

    I note that you did use two different versions of NINA. 

    So what I can say is that your dark does not characterize the light frame. You can blink the raw dark and raw light and see how the hot pixels are different. That is the bottom line. You need to take new darks.

    But there is one other thing- even with this issue- if you have well dithered data actually the integrated masterlight will likely turn out just fine. You might not be able to generate a dark that matches the lights if things have changed since you took the lights. I cannot tell the chronology of events.

    -the Blockhead
  • Still_Cloudy
    Hi,

    I totally agree that the darks aren't a perfect match, which is surprising to me.  I had also blinked the raw dark and the raw light and noticed that many, but not all of the hot pixels were a match in both images.  i had thought that darks were good for a long time, certainly months; obviously I will no longer assume that.

    The creation of these subs was my homework for the process of using ... for the first time ... NINA, PHD2 and the QHY600M camera, as well as a new guide camera (ZWO 174mini).  I'll certainly be much more diligent on my next target

    A thought: Could the data transfer introduce an issue?  I ask since the data transfer is complex, going from a USB hub to a converter to Cat6, to fiber optic to a remote computer.  That system is being replaced very soon with a computer at the scope which will communicate with another computer in my home.

    Thank you for taking the time to review these subs.  I may not be able to fix this image, but I'll not likely have this same issue again.

    The FastTrack course has been excellent.  I've gone through it thoroughly once with the course material, and again with my own data, and it has been immensely helpful.  I'm no longer afraid to attempt using Pixinsight.

    All the best,

    Bob H.


  • BlockHeadBlockHead
    Data transfer... it could, but unlikely. It is generally either going to work or fail with corruption. You don't get bad data values scattered around in this sense.

    -the Blockhead
  • Still_Cloudy
    Hello,

    Thanks again for the insight.

    Bob H.
Sign In or Register to comment.