Pau Gargallo
Pau Gargallo
That is a good point. Also, having this in a separate repo is not very practical. I'll try to move everything to the main OpenSfM repo. I'll revert this repo...
The above error message seems to be from the old `Interface.h` version in OpenSfM, not the newer one from openMVS. Are you using a 32-bit system? `size_t` and `uint32_t` seem...
That does not look normal. The problem does not seem related to openMVS though, so it is better to track it on the OpenSfM repository. I've just added an issue...
aha, I see thanks. That would explain the original problem where file exported by opensfm did not load in openmvs. I guess that keeping `Interface.h` updated is still a good...
I think that it is reasonable to ignore tiny focal lengths since they are unlikely to be real. It is difficult though to know where to put the threshold. A...
@russkel, I've resolved the conflicts on this branch and put the result in this branch https://github.com/paulinus/OpenSfM/tree/sashichandra-dsift About merging it to master, the main concern is that Dense SIFT is not...
@Juuustin, for the documentation, the error you pasted says that it can not find `sphinx-autobuild`. Maybe that is missing on your system. For the viewer, browsing http://localhost:8000/viewer/reconstruction.html#file=/data/berlin/reconstruction.meshed.json should load the...
Tests fail because `psutil` is not in `requirements.txt` @kjkim-kr, did you manage to run opensfm successfully after this fix? If so we could consider adding the dependency.
That is really great @BrookRoberts ! i need some time to parse it. The tricky part is updating the tracks graph. I wonder if it is possible to do without...
What should we do if one image does not has GPS and match selection using GPS distance is enabled? The current behavior is to deactivate the GPS based match selection...