lbestock
commented
4 years ago Discussed briefly today the need to store folders of images that have not yet been processed into models but need to retain their relationship with one another in order to one day be so processed. They have zero point as individual images; they matter only because they can be processed collectively to form a model. But one doesn't just do that, one waits until one needs something from that model. See also Alex's response about Pompeii and how they deal with gaps in the recording, #31. They, too, use models extensively, even to the point of ignoring elevation in excavation because it can be recovered from the models. To get elevation data from the model requires GIS. But to use the file repository to store and then find the files to produce the model of a context is something we will want to support, I am really pretty sure. Those files need to not only not go back into the recording system, they also need to not clutter an image search for a context in the file repository. This is why I think of a folder "Modeling shots Locus FA-001, Opening" or some such. Searching FA-001 then gives you all the locus photos you expect from the recording system and the QR codes, but rather than showing you the individual shots to be turned into a model you just see the folder on a search. One would need, too, then, to be able to download them in bulk rather than individually.
This is in fact a somewhat separate issue from how to deal with models themselves once they are produced, except that models cannot exist without reference to their photos and file paths on this front have always been a nightmare. How would a folder in the file repository (something we have not even discussed previously so I don't know if it can exist!) with the images to model interact with a model once made? I was thinking of downloading the images and then producing the model. Are there other options (and really, ones that do not burn the entire Amazon every time a team member wants to look at the model?)
Candace and Tyler, can I ask you please also to make notes here on anything that Connor comes up with regarding how to access Brown's roving license for Photoscan, and what workflow is for model making given that software usage here?
urapadmin
How can we store 3D models? For USTP they are a major part of the recording. Can they be stored in a useable way without storing the photos they are made from? But in fact, don't we want to store the photos they are made from? We do NOT want those photos to end up anywhere in the recording system, nor do we want them to come up in FR image searches. But if the FR is to hold all the project's files... But huge.