Tiled Model Creation and Editing: Impact of Depth Maps and Gradual selection

[314]

Hi Alexey and everyone,
I need some advice, please. I am creating a large tiled model from depth maps. My workflow was:
"Align Photos" → "GCP" → "Update Transform" → "Gradual Selection" → "Clip Cloud" → "Optimize Cameras" → "Tiled Model" with the option to recompute depth maps.
The problem is that the tiled model includes clouds that I had removed earlier in the workflow (specifically, I removed the tie points). Does this mean that editing tie points with Gradual Selection (and optimizing cameras?) has no impact on the final tiled model generated from the newly computed depth maps?
Is there a way to edit the tiled model? Can key points, which are used for depth map calculations, be edited?
This project involves a large industrial site with complex objects and structures, and I am limited by computational resources. I tested a tiled model generated from a dense cloud, but it is not as geometrically accurate. A classic mesh model with textures would be too large in terms of data size. Masking the photographs is also not feasible because there are 23,000 images.
Ideally, I need to remove the clouds (on sky) before exporting to *.slpk or find another solution.
Thank you very much for your help!

[JM3D]

I second this question...
If I could remove tie points that are obvious vehicles, it would be great if they could be gone and never be heard from again. I can fill the holes later after the mesh is created.