Debugging star detection during star alignment

[VoidPointer]

Hi,

I have a set of calibrated images where I am struggling with registration.
The process reports successful registration but the images are registered incorrectly. Basically, for all images it reports dx/dy values of 0 (or near zero like 0.1 px) which must be wrong since the images are dithered and dx/dy values of around +/- 10-50 px would be expected.
I guess noise and or hotpixels are being mistaken for stars.
By playing around with detection/noise scales, noise reduction, sensitivity and peak-response I got the images to register eventually but it was quite a lot of trial and error because if I changed a parameter in the wrong way, some images would just fail to register (not enough matching stars found)

Is there a way to determine which stars are being detected given the parameters that I enter? The images are quite noise but visually looking at an autostretched version of them it is pretty easy to tell apart noise and stars.

Attached is a screenshot showing one of the frames (autostretched) and a 1:1 magnification of an area with a few stars

Thanks,
Lars

[pfile]

yes - StarAligment has a mode called "detected stars" - it will generate a copy of the target image with crosshairs shown on every detected star.

rob

[VoidPointer]

Awesome, that's pretty much exactly what I was hoping for. Just upping the noise scale and tweaking sensitivity stops the noise being detected. Great to no longer be flying blind :-)

Thanks!

-Lars