Hi David,
We are doing the same as you are, only that we are using an RGB (GoPro) camera for the non-geotagged images. We have had no problems in getting both the geo-tagged Sony a6000 and the GoPro cameras aligned in the same alignment batch and judging by the sub-pixel alignment accuracy of the GoPro images, don't seem to suffer any degradation in overall accuracy. We still have to evaluate the resulting point cloud by means of validation points.
However, if in a separate alignment we use only the GoPro images (now geotagged with the positions as estimated in the first alignment), we strangely get large discrepancies between alignment output and geotagged camera positions. This does not make sense to me at all. Until we have an explanation for this unexpected discrepancy we will have to always include the geotagged imagery (as baggage) in the approach. Nevertheless, the process does seem to give us the ability to align secondary (non event marking) cameras in combination with V-Map geo-tagged images - thus obviating the need for ground control.
We'll keep on working towards a more efficient solution.
Walter