Components merging

Hi everyone,

I started my first project in Reality Capture - photogrammety mapping with my Dji Phantom4 ProV2 of an old historical building.

After importing and aligning images, SW creted two components - each containing 2 sides of a building.
I tried to merge them into a single model by using control points (Aligment/Merge Components Function). Although I added 10 points, the SW still fails to combine the images into a single model. Since I am very new to Reality Capture, but have deep photogrammetry knowledge and I am quite experienced in other photogrammetry SW (Bentley ContextCapture, Pix4D, Metashape, etc) I am wondering what I did wrong here, especially when both Metashape and ContextCapture managed to do the stitching and generate the model without any tweaks.

It seems like RC split components based on how I did flights in the field - First, I mapped nadir and one side of a building (image RC1) then I landed the drone, switched the batteries, and did 2nd flight, where I mapped the other two sides (image RC2)

It would be amazing if someone could help and clarify on what went wrong here and how to fix the issue.

Best regards,
ZoMi

Hi @ZoMi1987

I can see you are still using RealityCapture. You can try to use the latest RealityScan release, where are some improvements in the feature detection (which could help you in the alignment).

As I see the components, the Component 0 seems to have wrong GPS information. Could you select all images from this component and turn off their camera priors (in 1Ds view, under Prior pose/Absolute pose - set it to Unknown) and align again.
If you don’t need the georeferenced model you can turn off that for all your images.

After import of the images, are they placed properly in the Map view?

Hi,

Thanks for the answer, with RealityScan it worked! Let me continue with processing to see what is the final result

Best regards,

ZoMi