Hi,
I downloaded the trial version and am finding it very easy to use, but where can I find help topics on advanced features and settings?
More importantly, I have 18 images from a body scanner that I’d like to align and generate a mesh for. Using the hardware vendor’s software I can generate a full body OBJ, but using Reality Capture I can only get the front of the person. I have 8 images on the back that are not aligning and am wondering if there are settings I can change to add those to the alignment (hence my question about the help). Starting from the default settings, what settings might I change to help improve the alignment?
Thanks!
Hi Steve Park
18 cameras for full-body (FB) scanning is not a very good solution… It is a non-moving FB setup or the camera - subject is rotating and you have more than 18 images per body scan ??
The scanner is a structured light scanner that generates very accurate 3d body meshes from a static images. The subject does not rotate.
I am interested in seeing how well the RealityCapture software can reconstruct the mesh from the 18 images.
It seems like I need to put multiple control points in the images?
Hi Steve Park
If it’s a structured light scanner setup, then you cannot expect good results from pure photogrammetry solution with just 18 cams.
You can use control point workflow but I can say right now the results will be very poor… For proper photogrammetry build you need at least 64 cameras to get good results…
If it’s possible, use carton boxes with newspaper on them to get better alignment results, then use the XMPs to transfer alignment to other datasets + set proper scale.