Mosaic backgrounds

[John_Gill]

Hi,

I am busy working on a large mosaic. Due to various conditions the backgrounds for each night vary quite a lot.  How can I get the "value" of the background and copy it to the other images?

Any ideas or suggestions would be most welcome.

Look up
John

[pfile]

there are many ways to go about this. first is "frame adaptation" in staralignment when you are building a mosaic pane-by-pane. SA does a linear fit between the two panes before merging them.

always be sure to do DBE to each of your panes to try to remove any gradients before doing anything, because they will interfere with any equalization method.

most of the current "best practice" mosaic methods use GradientsMergeMosaic to do the actual mosaic merge. to a certain extent, GMM can handle the differing background values in the panes.

David Ault wrote a script called dnaLinearFit which computes the intersection between two images and then applies a linear fit to one of them based on the differences between the overlapping parts. if GMM does not do the right thing for you, you can use dnaLinearFit to fit try to equalize all your panels first, then proceed with the projections needed for GMM (either using MosaicByCoordinates or StarAlignment).

rob

[John_Gill]

Hi Rob,

Thanks for your input.  I sometimes use Frame Adaption, depending on the results. In GMM, I use various setting for file order, overlay and average.  The biggest issue is each row of the mosaic was shot on different nights over 3 months. I crop, abe, histo stretch, mlt, mask, tvg, histo, curves.  Then on to SA and GMM. Each row of images (same night shooting) mosaics nicely.  But, when adding the images from next row, the backgrounds are too different.

I will try your suggestion with DNALinearFit.  Thanks again!
John