each with each vs structure from motion

it is possible to add a each with each alignment algorithm? My results are combined between separate components.

Hi Przemek

it is possible to add a each with each alignment algorithm?
you mean every image to try align to every other one ? we do it but quite differently as other solutions

My results are combined between separate components.
Can post some screenshots so can better understand your issues ? as there can be more workflow solutions to solve this.

there is actually a possibility to define which features are used when connecting components.
for example if you have common part of stairs with floor reconstructions you may want to mege components using overlaps which is fastest option.
on the other hand if you are looking through the windows i’d use all image features to get the best precision