How to texture imported scene from Reality Capture

[Hildon]

Hello Everyone,

I'm trying to import scene and mesh from RealityCapture, to take advantage of the great texturing capabilities of Metashape. The scene is exported from RC in bundler.out negative z format. I added photos, imported images and mesh per recommended there. However when I textured the mesh it turned out not so good, see image attached.

Any suggestion how this could be fixed on my side, or any hope to get a patch for this ?

thank you in advance

[Alexey Pasumansky]

Hello Hildon,

Have you checked, if the cameras are positioned properly relatively to the imported mesh? For example, if you use Look Through command for some cameras, is the Model viewpoint orientated according to the corresponding image content?

[Hildon]

Hello Alexey,

Thanks for the suggestion. I checked and as you said some images were not aligned. The alignment and meshing were done in RC and exported to Metashape though. Will you kindly give me some advise about what should be done to fix the issue?

[Alexey Pasumansky]

Hello Hildon,

If you are using bundler.out format, it should be accompanied with the image list file with the proper order of the images related to the cameras in Bundler file. Also you should have the same set of images added to Metashape project that has been used for the external alignment.

[Hildon]

Hi Alexey,

I imported image list and the alignment became more accurate. However when I imported and textured the mesh the result was still not so good. I have attached an image for reference. Will you kindly give further advice?

[Alexey Pasumansky]

Hello Hildon,

If it is possible, please send to support@agisoft.com the original mesh model, bundler.out file, image list file and Metashape project file (in PSZ format) with the loaded thumbnails (project should be saved after all the thumbnails are loaded).
if you can also provide a few original images it would be even more helpful for our internal testing.

[Alexey Pasumansky]

Hello Hildon,

The provided bundler file includes only focal length information for the cameras without any distortion coefficients. Does it mean, that the original images are distortion free, or whether the distortions have been converted in RC, and it is expected, that now undistorted images should be used for texturing?
Do you have any information regarding adjusted interior orientation parameters (calibration coefficients) in RC after the camera alignment?

[Hildon]

Hello Alexey,

the RC alignment used  K + Brown 4 with tangential 2 distortion model setting. Will it be more helpful to you if I can provide RC processing files?