Official comment
I can't find an account for your name or email address so we can't look it up.
In general, you should only process images that are taken from a single takeoff location into a single map. If you mix takeoff locations and are providing ground reference images the last one encountered in the dataset will be used. Also, if you are caring about the accuracy of the elevation numbers in your results you really need to be using 80% overlap for you data collection. You can get away with less if you are just looking for a nice ortho image but elevation accuracy really needs at least 80% overlap.