Incremental mesh from depth maps

[BjFaeTorphins]

Hi Alexey
I have a project with 16,000 aligned cameras. In order to get a quick(er) view of what the final model would look like I disabled three of every four cameras then ran the mesh from depth maps.  So now I have a colourised, shaded model made from 4,000 depth maps.  I can see some holes in the model so what I tried was to enable more cameras in the areas of the model where there are holes then I re-ran the mesh from depth maps checking the option to keep depth maps. 

I had hoped that this would incrementally generate depth maps for the images that I had enabled (in the same way that you can add extra images to the alignment) while keeping the depth maps it had from the first iteration.  It did not seem to do that, instead it went directly to the building the model stage. I stopped the process as soon as I saw it going directly to build model stage.  Can you confirm that I need to rebuild all the depth maps?  If it does not do that could you consider it for an enhancement to a future version?

[Alexey Pasumansky]

Hello Brian,

Currently I don't think such kind of incremental depth maps generation would work properly, since if you enable the cameras that are overlapping with the original images used for initial depth maps generation, it would be necessary to re-calculate the original depth maps as well, otherwise, they would be based on incomplete information.