We are sorry to hear you aren't happy with your results. It looks like there were two separate adjacent missions involved here. Combining data from separate flights is always hard due to changes in lighting and shadows. We recommend having a couple of passes worth of overlap between them (not just having them abutting) and trying to match the lighting conditions as well as possible.
The bigger issues is that the overlap report for the top section shows that the overlap was marginal at best, likely around 65%. Basically you got lucky that the top section worked on its own but when used in conjunction with other higher overlap data it wasn't consistent enough to be included.