Hello, I am generating a photo set from a camera rig. I’ve exported a Reconstruction Region and use a custom CLI script to run through the asset creation, making an obj and a glb. However, I am having to resize the reconstruction region manually for every picture set, even though the rig has not moved and the subject is basically in the same place. This is extremely frustrating as I’m trying to automate asset creation from this, and the only thing wrong seems to be the fact that the reconstruction region wont stay centered around the point. How can I make it consistent between photosets, since the photos should be pretty standard each time? Thank you!
Hi Trent, the most likely cause is that you don’t have a defined coordinate system and every project is calculated in an arbitrary coordinate system. It has a different scale, different rotation, etc, even when the object appears in the center of the grid.
The best way would be to define a local coordinate system with a few coded markers for your rig. If your rig is solid and the cameras don’t move, you can reuse the existing camera poses as priors for the next calculations. This will preserve the local coordinate system. In this case, the reconstruction region will always be placed in the same position.