NormalizeScaleGradient: Bookmark website now!

Status
Not open for further replies.
The negative value truncation is always due to cold pixels. Truncating it to 0.0 actually helps because ImageIntegration will ignore all zero pixels.

The next version will only output warnings for high truncations.

1621425175286.png


Fixing the high truncations is easy. In this example use PixelMath to divide the reference frame by 1.1610 and then run the script with the new reference and the original target images.

Tip: If you want to quickly check for truncations, set the gradient smoothness to maximum (4.0). The script then runs much quicker.

Version 0.11 is attached to message #47

NormalizeScaleGradient: Bookmark website now!

The negative value truncation is always due to cold pixels. Truncating it to 0.0 actually helps because ImageIntegration will ignore all zero pixels. The next version will only output warnings for high truncations. Fixing the high truncations is easy. In this example use PixelMath to divide...
pixinsight.com
 
Last edited:
Thanks for the process!! I'm running it now on some data to see how it integrates into my workflow.

I'm curious about drizzling the files after they've been normalized and stacked.
 
The script works for a little while, and then I get error that reads "Normalized Scale Gradient is not a function" and refers to line# 91 in the script. If I quit PI and restart that sometimes fixes the issue, but not always.....
 
dcm_guitar said:
The script works for a little while, and then I get error that reads "Normalized Scale Gradient is not a function" and refers to line# 91 in the script. If I quit PI and restart that sometimes fixes the issue, but not always.....
Click to expand...
I am on holiday at the moment. I will look into it at the weekend.
 
dcm_guitar said:
The script works for a little while, and then I get error that reads "Normalized Scale Gradient is not a function" and refers to line# 91 in the script. If I quit PI and restart that sometimes fixes the issue, but not always.....
Click to expand...
The bug was in an error dialog that should have been displayed if there were less than three samples. Three or more samples are required to produce the surface spline model of the relative gradient. The bug prevented the error dialog from displaying. The attached version should fix this, along with a few other minor improvements.

Regards, John Murphy
 
If an image contained too many cold pixels that were black, too many samples could be rejected. I now allow more cold pixels before a sample is rejected.
 
The new PixInsight release 1.8.8-8 contains NormalizeScaleGradient v1.0
The help documentation is now included.
 
I have 1.1.8-7 (and manually loaded the script). I did a check for updates under the resources menu and it told me my installation has no updates.
 
you can't install major updates to PI using the update system. you need to download the new installer package, uninstall the old version of PI, and then install the new version. on macosx this is easy, you just delete or rename the PixInsight folder in /Applications and then install the new one. on windows and linux it's a little harder. the installer for the current version you have installed on windows should have the option to uninstall the software.

rob
 
jmurphy said:
I used the following PixelMath expression to find it:

iif(M101_-10C_600s_G10Off70-0080IR_c_cc_d_r == 0, 0, iif($T == 0, 1, $T))

('$T' target image is the normalized M101_-10C_600s_G10Off70-0080IR_c_cc_d_r_nsg file). Undo / Redo then shows flashing dark to white pixels.
Click to expand...

Hi John,
We were looking for truncated negative pixels.
I copied the above expression to PixelMath and get the below error when I apply it to the .nsg image
Error: Invalid character in numeric literal.

I cannot find any obvious typos, and I am a PixelMath beginner. Any ideas?

Thanks,
Roger
 
rockenrog said:
Hi John,
We were looking for truncated negative pixels.
I copied the above expression to PixelMath and get the below error when I apply it to the .nsg image
Error: Invalid character in numeric literal.

I cannot find any obvious typos, and I am a PixelMath beginner. Any ideas?

Thanks,
Roger
Click to expand...
I made a mistake when quickly typing the PixelMath expression. The problem is that the filename contained '-' character. PixelMath uses the display name, which drops the '-' characters. Personally I avoid using these characters in filenames to avoid this type of problem.

The PixelMath expression should have been:

iif(M101_10C_600s_G10Off70_0080IR_c_cc_d_r == 0, 0, iif($T == 0, 1, $T))

Perhaps a better example would be:

iif( myImage == 0, 0, iif( myImage_nsg == 0, 1, myImage_nsg ))

or

iif( myImage == 0, 0, iif( $T == 0, 1, $T ))

$T represents the target image
If myImage pixel is black, the target image pixel is set to black.
Else if the target image pixel is black, the target image pixel is set to white.
Else the target image pixel is left unmodified.

Hope this helps
John Murphy
 
Last edited:
Hi John .... Thanks for your reply.
I only upgraded PixInsight late last year and can't afford the latest version at the minute.
I was hoping that your script might still be available to download separately so that I could use it :-(

Cheers
Mike
 
MikePaling said:
Hi John .... Thanks for your reply.
I only upgraded PixInsight late last year and can't afford the latest version at the minute.
I was hoping that your script might still be available to download separately so that I could use it :-(

Cheers
Mike
Click to expand...
For major PixInsight updates, you need to log into the software download area. So far, all these updates (including this one) have been free for everyone that has purchased PixInsight.
 
Hello, since the script runs on already registered images, how can I drizzle after using the script? ImageIntegration does not let me load drizzle files
 
Last edited:
jmurphy said:
For major PixInsight updates, you need to log into the software download area. So far, all these updates (including this one) have been free for everyone that has purchased PixInsight.
Click to expand...
My current version of PixInsight is: "v 1.8.8-4" ..... it tells me that if I need to access "v 1.8.8-8" I need to purchase a whole new copy before I can try out your new script.
Am I missing something here?
Mike
 
MikePaling said:
My current version of PixInsight is: "v 1.8.8-4" ..... it tells me that if I need to access "v 1.8.8-8" I need to purchase a whole new copy before I can try out your new script.
Am I missing something here?
Mike
Click to expand...
Take a look at the license information in this post:

PixInsight 1.8.6 Ripley Released

I am happy to announce that PixInsight 1.8.6 has just been released. This is a very important release with many new features and performance improvements, with a strong focus on stability and compatibility, and a much increased potential for future development. Installation packages are now...
pixinsight.com

If you still have problems, I would recommend contacting Juan Conejero .
 
Last edited:
Status
Not open for further replies.
Back
Top