Error when building tiled model from a loaded mesh

[rubmart]

In one of my work pipelines I use Metashape to build tessellated and textured models from existing 3d models by loading the solid model. We load the solid model in STL format inside Metashape to texture it and build the tessellated model.

Original model is a STL file, see imagen model_solid attached to the post.

This process worked correctly in version 1.7, but now in the new version 1.8.1 and tested also in version 1.8.3 the tessellated model destroys the original geometry. Instead of triangulating correctly based on the solid model, it destroys the geometry and forms new polygons where they did not exist before.

Resulting tessellated model (version 1.8.3), see Tiled_Model_Solid image attached to the post.

Any idea why this may occur in the new version?
(in version 1.7.6 we have normally worked with this kind of formats and conversions without problems, we have tested the same model in this version and the tessellated model is created correctly).

[Alexey Pasumansky]

Hello rubmart,

Can you please send any example of the external model to support@agisoft.com, that we can import to Metashape and use as a source for the mesh generation procedure in order to reproduce the issue?

Also please specify, if you are observing the issue on every zoom level of the tiled model, not only on the less detailed.

[rubmart]

Ok, no problem, i will send via email more info about the problem.
Thanks!

[Alexey Pasumansky]

Hello rubmart,

Thank you for sharing the sample file.

The problem may be related to the mesh vertex coordinates rounding issue. The imported model has vertex x coordinates in 450823.0 - 451174.8125 range and y coordinate range in 4480485.5 - 4480754.0 limits. If you are importing mesh model to the new empty chunk, I can suggest to do it using the following Python code line (input to the Console pane) that would subtract the high base values from X and Y coordinates:

Code: [Select]

Metashape.app.document.chunk.importModel(path, shift=Metashape.Vector([-450000, -4480000, 0])) #values specific for the provided model