I've been using BPP for a while now and not had any issues with it working until I changed my camera from an Atik 11000 to a Moravian G4 16000. Since then all my attempts to use the BPP end in failure with the image registration and RANSAC error failure to find star matches. I've tried some suggested fixes including all the pixel interpolation options and upping noise reduction but all with the same outcome.
However, if I take the calibrated images from the BPP instance and use staralignment - they all register without any issues?
I found this much older thread
https://pixinsight.com/forum/index.php?topic=4636.15which claims that the error I am currently encountering was fixed in release 1.31 of BPP. If this was the case then it seems to have returned as my experience is exactly the same as people were having back in 2013. None of the images I am processing have been flipped so that is not an underlying issue with my data either.
Can anyone offer any help?
UPDATEI received some advice from Pixinsight Jedi Grand Master pfile who suggested:
usually a ransac error occurs when there are a lot of hot pixels or other features that the star detector has erroneously identified as stars, and then staralignment is overwhelmed and cant find star matches between the reference and target image and if your darks match the lights in duration and temperature then it might be worth turning off dark optimization during calibration and see what the calibrated files look like wrt hot pixels (if that even is the problem.) or, you could try CosmeticCorrection to clean up the hot pixels again if they really exist.I did some tests and tried a mix of these suggestions and found one that worked. The solution for me was to reduce the default setting of max stars from 500 to something under 100, I'm using 80 and that's working quite reliably at the moment.
Thanks to Rob for saving my sanity as I just couldn't find out what was wrong - I think I tried everything except what worked in the end.
