I’ve been wondering about the effect of moiré patterns in images when doing photogrammetry.
Is it a common problem in certain areas?
When doing ordinary photography it can be a problem, fine patterns can look like crap.
2 Likes
I think things get smoothly resampled enough in the texturing process which itself is relatively stochastic and patchy that any existing moiré gets masked in photogrammetry. Total hunch here: I have yet to see it in drone photogrammetry.
You’ll note that the texturing is lossy: it’s never quite the detail that went in, and this is probably a large part of the reason we never see it.
2 Likes
What about the point cloud?
1 Like
I’ve never once seen it on even with data from my full-spectrum converted cameras (which lack OLPFs which tend to mask moire), so I don’t think it’d be a real issue. Most of us aren’t doing reconstructions of textiles or LCD panels at a distance where moire would affect us.
Moire tends to only happen in the recorded images when the pattern aligns with the structure of the photosites on the sensor in a particular way… It really is hard to have happen. Preview images on the rear LCD of a camera may show heavy moire, but usually that’s just really bad resampling of the preview image (see below).
If anything, you might see it in the pyramids/overviews of the export GeoTIFF if the resample method isn’t appropriate, but it shouldn’t be “real” in the final products.
2 Likes