Avoid cropping when combining sets of images taken at different times

[ram]

In PixInsight, when I integrate lights from different days or especially if I move things around (rotate the camera, etc.), even though I try my best to nail down the framing and match the frames, etc, I am off by a few to several pixels which is noticeable in terms of the edges being "visible" in most cases. I've gotten much better at this over time but still, let's assume the goal is to make almost every pixel count (anything in excess of 5-10 or so pixels on each edge). Let's say we're talking a change of 100 pixels due to framing after everything has been cropped to remove the black edges and we're talking about a single filter mono image (since if we went with multiple filters the problem actually can get worse). (Or in this case today I had to change the spacing between the reducer and sensor and reattach the camera so that changed the framing slightly and I don't want to throw out the 3.5 hours of data I already collected.)

Now previously I did a mosaic image and I was able to use DNALinearFit to equalise the images and then used gradient merge mosaic to combine them but I wonder if that would work for this (and it seems it would just the pairs of images to be merged would be very very overlapping) or if there's another strategy that works better.

Thanks all!

--Ram

[yobbo89]

try star alighn your images with thin plate splines - distortion correction/local distortion and with frame adaption on .

[STEVE333]

I use APT to control my equipment during data acquisition. It allow me to use a Plate Solve routine to accurately center the desired aim point. This means that every time I go out to capture data for a given target I can accurately center my FOV to the same spot. This doesn't correct any rotation misalignment, but, corrects everything else.

If data is collected in this manner, then, use StarAlignment (yobbo's suggestion) to correct for any remaining translation/rotation errors. Not sure you can do any better than this.

Steve

[Star Dad]

Nebulosity can correct rotation. When aligning stars for stacking it requires that you align two different stars in order to calculate the roation. I've done this with images taken months apart. Yeah, you lose some peripheral areas, but that shouldn't matter. Here's a link: https://www.stark-labs.com/