Camera selection groups for reuse with processing wizards and tools.

[polarathene]

Sometimes it would be nice to only use a subset of cameras for processing steps(or tools like photoconsistency). This can help reduce processing time or memory usage(as is the case for me that 50 8MP images at 100% and one 8k map often fails during the process step with 8GB RAM(I have 8GB virtual memory set aside on disk but Zephyr seems to consider that over 100% memory usage).

Recently I was using this for the texturing phase and out of 140 photos I used to reconstruct with earlier phases, producing an 8k texture would fail unless I used about 40 images. No problem, the wizard allows me to select individual images(from thumbnails that aren't exactly easy to evaluate quality). Problem is that this list is temporary, if the process fails, or I want to try different settings(weighting, number of textures or size of textures, etc), I must manually go through scrolling the thumbnails and selecting again(which is really frustrating if you accidentally deselect as you lose the whole selection and have to start again).

I would love it if I could save the list of cameras I selected/filtered(eg I could create such a list with the Image Quality Index utility, instead of using that selection for a new project, sometimes I want to optimize/reduce my cameras but it's difficult to tell which are most appropriate, additionally one could run such a filter utility on an existing list/group of cameras in the project). These lists can be reused or modified(add/remove cameras). Perhaps you could create such lists with the existing UI for selection or the dual list for matching new photos to existing cameras, but include the ability to view/zoom larger version of an image(external image viewers are not always practical if you're already using a subset of a dataset from a directory of many images). Selection of lists can be a dropdown combobox, like you would select a dense or mesh element in the wizards.

[polarathene]

Slightly different feature request(could post new thread but would depend upon this feature).

Could an optimal coverage utility be made available? Where you can filter out camera's that provide less information/points and/or confidence in a points accuracy(weight?). It would work similar to the bluriness utility(and could additionally be combined with either utilities output camera selection group as an input), a slider widget to adjust how many cameras would be selected. Another parameter should be prioritizing coverage overall, so that point confidence isn't the only weighted value assessed.

It'd also be nice if you could run this utility on a selection of points/vertices(since there is already a feature that filters cameras based on selection, once this threads feature is developed, that should be doable).

This feature would be for filtering out(and later adding to a camera selection group if desired where needed) cameras from computation where they are not as necessary, speeding up computation time and reducing memory required to process.