preferred method to align outputs from multiple cameras

I fly an ag field several times throughout the growing season using a Mavic 3 Multispectral and Mavic 3 Thermal. I’m also using a D-RTK 2 on all flights. This produces 3 sets of data - rgb, multispec and thermal images. I’m currently processing the 3 sets of images separately, Pix4Dmapper steps 1 - 3. Individually, the outputs look as expected. When performing post-processing in GIS software using all Pix4D output orthomosaics/indices I can see that there is a slight shift between the geo positions of various images. This skew between images is causing issues. What is plant material in one index, is soil in another, etc…

Reading more, I found others recommended creating a merged project. I created 3 sub-projects and ran only step 1 processing. I added 5 MTPs (field markers used to assist with GIS processing) to each sub-project, manually marking several images in each band, including each of the 4 multispec bands. I created a merged project using the 3 sub-projects. All 6 cameras were detected, but only rgb indices were available to select in step 3. Not sure what happened to the multispec and thermal indices, but I ran it anyway to see how the outputs looked, but it failed during step 3 processing…

I’m now trying to decide if it is worth troubleshooting the merged processing. If I can get it working, will it provide indices for all bands? (3 rgb, 4 multispec, and 1 thermal) And most importantly, will the images be spatially aligned with each other? Would I be able to accomplish the same thing by running each set of images as a separate project (steps 1 - 3, which works well, no errors) and using my existing, anchored field markers as GCPs? I haven’t done so, but imagine I could get their positions using the D-RTK 2 or maybe even pull values from an existing orthomosaic. If I include the same GCPs in each project will that generate outputs with better alignment with each other?

This seems like a common problem, but I only do this type of work a few times a year. I’m hoping someone with more experience can point me in the right direction.