I have tried two different methods for georeferencing my point cloud and mesh. With neither method have I managed to produce something I can use.
I use georeferenced images, The images are georeferenced in WGS84, including pitch/yaw/roll. I want to constrain this model using high contrast markers (ground control points, which are georeferenced in British National Grid/OSGB36. I specified that the images are in WGS84, and that the ground control points are in OSGB36 - and furhtermore that I want the output model to be in OSGB36.
When I import this into a GIS software or CloudCompare I can see that the points cloud and mesh are not properly georeferenced. The point cloud and mesh are misaligned compared to other data products in OSGB36 coordinate system.
To solve this I tried using images that were not georeferenced. I only constrained the model using ground control points georeferenced in OSGB36. However, this produce a visually poor model with a severe bowl effect, and which seems to be stretched in one direction. I used 11 ground control points if that matters. The area is roughly 6000m2. I flew at 50m elevation.
I would appreciate any feedback and comments as I am new to Pix4D (although not new to Structure from Motion). Please let me know if you would like me to attach parts of the output.
Thank you in anticipation.