calculate new (higher point) point cloud/model from selection?

[James Beckett]

Hi there,
Struggling to find a good workflow/ tutorial on the task

I am looking to avoid laborious masking of landscape photos. For this purpose, is it possible to generate an initial point cloud, then make a selection and run a higher point count on that specific selection (as opposed to the full range of potential points from source images)? This selection should naturally hold through mesh and texture processing.

I can almost manage something similar, by generating a model, cropping, then import masks>Method: from model.  The masks are visible on each respective image, but when generating a new point cloud, the whole set is again sampled, resulting in a lower point count than desired. Naturally setting all settings to high would just result in longer processing times.

Any tutorials on the subject much appreciated (many ways to skin a cat I realise)
Big greetings and thanks in advance jjj

Version 1.6.2

[Costas]

Calculations take place for everything that is inside the Region Box. Thus by scaling and moving the Region Box you can limit the area of the calculations.

In your example you can:

1) Run initial alignment (High)
2) Scale and move the Region Box to the area you want. eg. so only the building is inside the box.
3) Continue to other steps to calculate Dense Cloud, 3D Model etc.

you can find the Region Box in the toolbar (photo attached)