cannot reproduce Astrometric Solution/Image Solver failing on PI 1.9.2

0

01blueglide

Well-known member
Updated to PI 1.9.2 on Intel MacBook Pro 16Gb. A couple of unknown intermittent crashes (one during WBPP) but second attempt WBPP ran without crashing on data (M1) I previously successfully processed on PI 1.8.9-3. WBPP processed without errors until astrometric solution step in the pipeline when PI reported it was unable to provide astrometric solution. Ran DBE, Blur Xterminator on master XSIF but Image solver failed. This same data successfully solved on older PI version. No changes in workflow or processing parameters. Tried a different set of data (NGC7635) that again was previously successfully processed on PI 1.8.9-3. Again, using identical workflow, no astrometric solution found. I have no idea what is going on. Going to roll back again to earlier version. Attached log but I don't know how to troubleshoot.
 

Attachments

There is no crash at all in the log file you have posted, just failures to calculate astrometric solutions. Can you please upload a data set or an image where this problem can be reproduced?
 
More useful to post one of the master frames that won’t solve.

Cosmetic Correction is finding 180k pixels to correct, which seems pretty high?
 
Last edited:
OK I'll try and upload to Google Drive. The image was pretty noisy prior to processing thru DBE, Blur Xterminator, etc. Afterwards, prior to SPCC I was able to successfully solve in Image Solver when using PI 1.8.9-3.
 
Thank you for uploading the image. It can be solved with ImageSolver 6.3.1 on PixInsight 1.9.2 on Linux, Windows 11, macOS 15.2 M1 and macOS 15.2 Intel without problems. Default ImageSolver parameters. Please check that you have Gaia DR3 database files correctly configured.
 
I'm stumped...I have tried solving this frame manually using default Image Solver settings...all Gaia catalogs are in same place as always and PI points correctly to them. I have tried solving after some noise reduction and background extraction...error. Tried solving frame with no post processing at all...error. I wonder if uninstalling and then reinstalling PI 1.9.2 would give different results. I get intermittent crashes with BlurXterminator so I tried solving frame without BXT. I can't solve using various combinations of workflow that normally worked fine. Think I'll try new install.
 
As an aside, our new arbitrary local distortion correction algorithm with DDM thin plate splines works wonderfully to correct lateral chromatic aberration problems. This is with your image, bottom right corner:

Desktop.jpg


Let: original image. Right: RGB channels splitted with ChannelExtraction, registered with StarAlignment, and merged with ChannelCombination.

StarAlignment with reference = green channel, two-dimensional surface splines, DDM thin plate spline, 4000 spline points (default parameters).
 
Juan Conejero said:
As an aside, our new arbitrary local distortion correction algorithm with DDM thin plate splines works wonderfully to correct lateral chromatic aberration problems. This is with your image, bottom right corner:

View attachment 25426

Let: original image. Right: RGB channels splitted with ChannelExtraction, registered with StarAlignment, and merged with ChannelCombination.

StarAlignment with reference = green channel, two-dimensional surface splines, DDM thin plate spline, 4000 spline points (default parameters).
Click to expand...
Hey, that's pretty slick.
 
Also double check that none of the database or working files are in a location subject to replication (Google drive or the like)
 
Screen shot of GAIA files. I am going to try re-processing data set again without Astrometric Solution in the pipeline. WBPP stalled at that stage and looped repeatedly before I could abort the process. I wonder if it corrupted somehow on my computer.
 

Attachments

01blueglide said:
Screen shot of GAIA files. I am going to try re-processing data set again without Astrometric Solution in the pipeline. WBPP stalled at that stage and looped repeatedly before I could abort the process. I wonder if it corrupted somehow on my computer.
Click to expand...
It looks ok.
Can you try to run ImageSolver on this image (the one you share in post #7) and share the complete console output?
 
01blueglide said:
I think this will work....first part of log was WBPP processing M1 data set again...I aborted process to run Image Solver on frame from post #7.
Click to expand...
There seem to be a lot of errors or warnings of the sort that are common when you're reading or writing files from/to folders that are being backed up to iCloud or managed by Time Machine or similar utilities.
 
cloudbait said:
There seem to be a lot of errors or warnings of the sort that are common when you're reading or writing files from/to folders that are being backed up to iCloud or managed by Time Machine or similar utilities.
Click to expand...
I dont use iCloud but I store my data sets on an external HD that is backed up to Time Machine regularly. I have been using this configuration for a long time. I create a folder on my desktop when processing in PI to put all the masters in and then transfer it to the hard drive later to save storage on my MacBook.
 
01blueglide said:
I dont use iCloud but I store my data sets on an external HD that is backed up to Time Machine regularly. I have been using this configuration for a long time. I create a folder on my desktop when processing in PI to put all the masters in and then transfer it to the hard drive later to save storage on my MacBook.
Click to expand...
Well, for some reason it looks like files are being created which then cease to exist. Typical for synced folders.
 
You must log in or register to reply here.
Back
Top