No need to worry about the yellow exclamation marks. The only cameras added to the database are the problem ones. You can also add cameras in a simple text editor if you want the yellow marks to just go away, but it has no bearing on the final quality of the model.
Also, from what I understand Reality Capture will down sample the real textures in the program to help speed up performance. If you want to see the true quality of the final output click “render” to get an image or export the model.
Thank you for information we will add sensor information to our database ASAP. Sensor information is helpful for the alignment but not necessary. Even without it RC can create proper alignment.
If you would like to you can add camera to database of your RC manually.
There is step by step guide in Reality Capture help. Please check “Extending Camera Database” section.
Regarding texture, RC sometimes does not display the full texture resolution in sweet render mode. RC is trying to fit them to GPU, max one 16K texture (or 4 8K textures, 16 4K textures, etc.). Therefore, the displayed texture can be at lower resolution than the calculated one. Is it blurry also when you export it?
Please could send us print screen of texturing settings used for the model?