OpenMVS with openMVG for the alignement works really very well!
The only problem is that for large blocks it is not able to terminate.
My computer has 16 GB Ram, 4 core 3Ghz Intel i5 processor.
For example I have a block of 200 16Mpixels (4920 x 3264) images. I use reduction level of 1 ( depth map calculated for 2460 x 1632) during the fusion step it exits with failure (catching an exception).
I try the same block with level 2 (1230 x 816) and in 40 minutes it produces an excellent 35 Mpoints cloud.
Is there a way to work even if the ram is not enought?
yes, there are ways to make OpenMVS support huge datasets, but are not implemented yet... this is on the enhancements list, and any help in coding is welcome...
OpenMVS with openMVG for the alignement works really very well! The only problem is that for large blocks it is not able to terminate. My computer has 16 GB Ram, 4 core 3Ghz Intel i5 processor. For example I have a block of 200 16Mpixels (4920 x 3264) images. I use reduction level of 1 ( depth map calculated for 2460 x 1632) during the fusion step it exits with failure (catching an exception). I try the same block with level 2 (1230 x 816) and in 40 minutes it produces an excellent 35 Mpoints cloud. Is there a way to work even if the ram is not enought?
Thank you Franco Flamigni (Italy)