StarAlignment not rotating to align

[heyjp]

I used StarAlignment and ImageIntegration pretty extensively last year (2014) on probably 50 or so image sets from my Canon 6D.  I put it away for several months and just downloaded the latest version (as of May 1 2015) for my MacBook Pro.  I am using Yosemite 10.10.4b3 and PI 1.08.03.1123.

My problem is that I have run several sets of images thru StarAlignment and the resulting "aligne" images have not been rotated and are not aligned.  I am shooting 30s subs on a Celestron CPC1100 wtith Alt/Az mount.  The resulting images definitely have the expected amount of rotation between images.

I still consider myself a newbie and still mostly a virgin. :-)  Here's my process:

 - I shot 50 Canon RAW subs.
 - Exported them as full size 16-bit Color TIFFs. 
 - Opened a new session of PixInsight and launched StarAlign
 - I added all 50 TIFFs to the target image list
 - I added the middle image #25 to be my reference image
 - I picked a folder called ALIGNED to place the resultant aligned images
 - Left everything else defaulted and hit Process Globally

It looks like it processed all 50 images in the console.  The alignment matrix looked funny though.  For most all images it showed small values for dx, dy but 0.00 degrees.

When looking through the 50 "aligned images" they still show signs of rotation remaining.  They are not aligned.

What stupid thing am I doing wrong?

Jim in Boulder

[Carlos Milovic]

Maybe it detected thermal noise as stars, so it was aligning on the hot pixels. Try to perform calibration first, and maybe use CosmeticCorrection or DefectMap to remove the worse hot pixels.
If the above fails, or not apply, change the parameter in SA to improve the detection. Change the output of the process to see the matched stars, etc.

[heyjp]

Thank you Carlos!  I will try your advice.

Jim in Boulder

[dsalman]

I had the same problem last night with 300 one shot color images using a Alt az mount (rotation)   There seems to not be a way to calibrate (color convert) images. 

[NGC7789]

dsalman,

I'm not so sure yours is the same problem. There's nothing to indicate alt az with the OP's problem. How long are your subs? Is it possible the star trailing from rotation is preventing identifying the stars? And what do you mean by color convert (I think I asked you the same question in your other thread)? If you mean debayering that's not the same as calibrating. If I'm not mistaken the sequence is to first calibrate, then debayer, then star align.

[dsalman]

I found the batch process.  The problem I think, is that the rotation is so much and the FOV changes over time that other stars come in as well.  For example, M82 was centered and M81 was out of the FOV but by the end of 3 hours M81 was in the field of view.  what surprised me is what Maxim can register ok.  But I suspect there is a tweak in the star alignment I need to do

[NGC7789]

I don't think change of FOV is the issue. Star alignment only needs a few hundred stars if that. Think about the mosaic mode. That is only using stars on the edge of the frame to align. There is some other cause for your issue.

[Cleon_Wells]

Jim, try using SA settings like these, I ‘m always tweaking SA settings. Try using a Preview from the Reference Image for a Reference area of the sub. When I’m chasing problems with SA I grab the first three, the middle three and the end three subs  to reduce the tweaking time.
This is M49 in the Virgo cluster, check it out, lots of galaxies.

http://www.atlasoftheuniverse.com/galgrps/vir.html

The Camera is a Canon  T1i at 400iso and 155sec guided subs.