chair without legs, why

[vicolodo]

Hello,

i like a lot 3dzephyr but sometimes i get results that discourage me

i use 3dzephyr free and my smartphone huawei p8 max, that has a 12mpix good quality camera

yesterday evening i did a try, taking 50 snapshots of a chair and processing them with 3dzephyr

i selected 'fast project' processing, with default settings
and after i retried with 'deep' sparse cloud creation and 'high details' dense point cloud creation

i thought to get good results because the chair is a static subject, and it was placed in a place with backgroud and ground rich of details to be get as reference points by the program

BUT the results were not good: whole legs, or part of legs were missing , and big holes were present in the final mesh:

https://1drv.ms/f/s!Ahoag6A7lek1jFmomkerVy6bBMu6

example dense point cloud:


example mesh:




This is the set of photos:

https://1drv.ms/f/s!Ahoag6A7lek1jB3ErBNFVP6eel7_

example photo:


As you can see in the final 3dzephyr screenshots, all the photos were matched, the program did not discard any of them,
but the results are not good.

why?

i wonder if it is mainly a camera problem or an illumination problem or a subjet problem
do you think that with a good canon dslr i could get better results?


thank you!

[cam3d]

Hi Vicolodo,

I believe it is simply that the plastic surface of the chair has very few features and as a result is difficult to reconstruct with such few images and a phone camera. Your images have all aligned perfectly because the floor beneath the chair is the IDEAL surface for image matching, it is covered in lots of lovely random features which make alignment much easier.

If you use a DSLR and a tripod you can dial in the camera settings to get optimal sharpness on the subject which will result in a far superior reconstruction

Setting suggestions: LOW iso (under 200), high F-stop (F/11-16), Shutterspeed variable if using a tripod (try and trigger camera remotely or with a 2second timer to avoid shaking the image accidently)

Hope this has helped somewhat! - Let me know if you have DSLR questions