I was hoping to export camera alignment and corresponding undistorted images from Reality Capture to Blender. My goal was to allow further mesh editing and texture projection in Blender. Unfortunately, I can’t seem to get this to work correctly.
After the reconstruction step succeeds, I select the reconstructed mesh and then go to RECONSTRUCTION > Export > Mesh. I choose the FBX format and use the following export settings:
The most important points are:
- Export cameras > True
- Fit > Inner region
- Resolution > Fit
- Undistort principal point > False
I then open Blender 2.80 and import the generated fbx file. I can see that the reconstructed mesh and its texture have been correctly exported, and that all aligned cameras are present and seem to have the correct position and orientation.
To notice the issue, I proceed as follows:
- select one camera
- go to Properties tab > Context > Background Images > Add Image, and open the corresponding undistorted image.
- in the viewport, select the “Wireframe” viewport shading method and click “Toggle the camera view”.
At this point, I can see that the background image and the wireframe mesh are slightly off, and that the camera object in Blender doesn’t seem to have the same aspect ratio as the undistorted image.
Does anybody have any recommendations as to how to properly setup this kind of workflow?