A tecnologia é simplesmente espetacular. Ela estabelece um novo patamar na captura e digitalização de ambientes para as mais variadas aplicações. O mundo dos games passará por uma grande transformação, permitindo a criação de jogos em cenários extremamente realistas, baseados em locais da vida real.

Hmm, I wasnt notified about your response and rather randomly saw it. However, Aligning 1776 photos in reality scan should be fast, but dont know if it gets the positions right. Masking blurry areas can be done with photoshop actions. I have done that already... the results on the one hand showed that this could work, but on the other hand, iirc I had only ~300 poses and after that it used only a fraction of that so the results showed bad coverage artifacts

@simon. Great question! That is a workflow I tried and might revisit. The issue is instead of aligned 111 photos, you'd have 1776 photos, even if you matched only relevant photos, that would still take very long for the SfM and some might not track at all. Masking the blurry areas is easier said than done. As far as I know there is no good algorithm that gives you the area in focus, doing this with some edge detect filters is not good enough. Any stacking software uses quite sophisticated algorithms to blend the images together. Your best bet would be to extract the masks from a stacking software.. but because they also align the images, it is also not very straight forward.

I was wondering if there is any chance to run a test with your source photos. I would like to know if it is possible to align the photos before focus stacking. And then mask the blurry areas before training.

I now learn more here on SuperSplat than at my university ))

DUDE