If a mesh is made from dense point cloud- what's the point of the sparse point cloud?

[SarahC]

Is the sparse point cloud just to calculate where the photographs where taken "in space", for the dense point cloud to be made?

I'm asking, because would it make sense to create a "Fast" sparse point cloud if possible to save time - and then a "Detailed dense point cloud" if all the photo's got included in the sparse point cloud pass?

[Andrea Alessi]

Hi Sarah,

yes, the most important output of the first phase is the actual camera calibrations and camera orientation rather than the actual sparse point cloud. For 3D modeling, you're probably not going to ever use the sparse point cloud but in some other cases some of our customers simply stop at the sparse point cloud.

It's absolutetly possible to use a faster preset for SfM and then more detailed presets for the other phases. Make sure however that the cameras are correct: even one incorrect camera can bring in a lot of noise.

On a side note, i have good news. Samantha (our SfM technology) is under a very significant speedup for the next version. You can read more on our previous status report here https://www.3dflow.net/3df-zephyr-de...gress-avocado/ and we'll talk more about speed improvements in our next status report

[SarahC]

I see, thank you for the tips Andrea.

The changes to Samantha sound very good!