How do you troubleshoot elevation discrepancies in UAV mapping model?

[archeomason]

Hello -
I am new to UAV mapping and have recently flown a mapping mission outside of San Antonio, Texas, USA that we processed through Metashape Professional. When I process the data and generate the DEM from the dense point cloud, it's showing up ~25 meters above the prevailing grade. It looks like a flat table in the "Model" window. It lines up perfectly laterally; just vertically it's off. As an example, the model's height is 157.595 m at the edge and just off of it (on Metashape's background imagery), the elevation is 132.242.

That leads me to conclude that I'm dealing with a datum/projection-related issue but I can't seem to find a foothold to pursue troubleshooting where the problem is stemming from.

We had some problems collecting the elevation values for our GCPs in the field. To compensate as much as possible, we used the USGS 3DEP Lidar dataset for the area to get a close approximation. The LiDAR dataset we used indicates that it's using WGS84, which is the same as Metashape's, but I'm clearly missing something. Is one using Height above ellipsoid but the other is using geoid? I don't know where to check and/or what to look for.

Has anyone dealt with an issue like this before? Can you make any suggestions on where I can search to figure out?

Thank you very much in advance for any help you may be able to provide.

[dpitman]

If your model's elevation is coming from  GCPs, then whatever datum was used for the GCPs will dictate the elevation of the model.  You would un-check the cameras (uas images) before running the Optimize Cameras command.

Many basemaps are using orthmetric elevation which is based upon a geoid. WGS84 is based on ellipsoid and in the San Antonio area, the web says that generally, the geoid height is approximately 23 m below the ellipsoid. That looks close to your stated discrepancy.