I have been working on making a reconstruction of a hanging plumbob and aluminum plate (see pictures). I placed targets on the plumbob and plate and I took a dataset of photos using flash, so the targets would light up. I got a pretty good reconstruction, but I wanted to try it without flash and see what happens. In my reconstruction with no flash, most of everything was reconstructed except for the plumbob (see picture). I am curious as to why the plate on the ground still showed up, but Zephyr had such a tough time reconstructing the hanging plumbob. I can see small parts of the plumbob and "shadows" of the it around the plate, but no 3D reconstruction. I'm wondering what specifically is confusing the software, is it becauause the plumbob doesn't have enough contrast on the surface (which is provided by the targets when lit up), is there a rule of thumb for how big the object needs to be in this situation (would a wider plumbob work)? Thank you!
Object Not Rendering with No Flash
Collapse
X
-
Tags: None
-
Hi mayaatassi - Without seeing the photos and knowing your capture settings, here are my thoughts:
Using flash generally increases the amount of light, which typically results in cleaner data and more successful reconstructions.
The dots aren’t doing much here. It's the surrounding environment that provides a grounding anchor for the subject. Background features help reinforce camera positions, leading to a more stable orientation around the subject. Without that context, the subject alone would be much harder to reconstruct.
To improve the capture, consider applying a fine light powder (such as talc) to the subject. This helps mattify shiny surfaces and adds more visual features, making it easier for the software to reconstruct even when camera alignment is solid.
Comment