Strange "Radial" DEM/Low Resolution Orthomosaic

[willcodeforfoo]

Sometimes, when processing historic air photos, we get a lower than expected resolution (e.g., 9m/px instead of 60cm) and a strange "radial" pattern in the DEM (see attached screenshot). Any ideas what might be causing this?



Thanks!
Kevin

[Alexey Pasumansky]

Hello Kevin,

How does the alignment look like (cameras and tie points) and what parameters for DEM generation you are using?

[willcodeforfoo]

Alexey: Thanks for the reply. Alignment seems OK. Attaching the report PDF, hopefully that answers your questions!

[Alexey Pasumansky]

Hello Kevin,

Are there fiducial marks on the images that you are working with? If so, then I would recommend to try re-aligning the dataset with fiducial marks assigned to the images.

In the report the side view of the alignment is not shown, but I assume that in the Model view you see the obvious bowling effect in the tie point view mode. Could be a result of distortion parameters estimation using Frame camera model, without considering the fiducial marks.

As for the default resolution value, the unexpected resolution is likely related to the DEM built based on the tie point cloud. If you use depth maps or point cloud based approach, you will get DEM resolution that relates to orthomosaic resolution by depth maps downscale factor.