Save alignment data when exporting masks to combine a top and bottom set?

Development Asset Creation
,
1

Lets say I put on an object on a table with markers to help with alignment, and i take a set of photos. Flip it and take another set.

When I reconstruct, the table itself serves a purpose during reconstruction, so it goes smoothly. But I have to export mask and cover it up it since I want to merge 2 sides.

When I import 2 sides now im getting not 1 whole piece but many smaller ones since ive lost my reference points from the table. Is there a way to save/export how the first set of images got aligned plus mask?

Ideal situation:

Side 1 > Align with the help of table with markers > export alignment from resulting model plus mask

Side 2 > Align with the help of table with markers > export alignment from resulting model plus mask

Merge 2 Sides > Import alignments and masks so that RealityCapture knows there are two models trying to merge here, and looks for reference points within those 2 models to try to piece it together. Worse case, it doesn’t match up and I have 2 parts to piece together with Control Points.

2

Hi SOLDIEROSCAR,
what kind of object and what kind of markers are you using? Does the table has its own features?
I would go like:

  • import side 1 set of images, align, create model and export masks
  • in new project import side 2 set of images, align, create and export masks
  • open new project, import masked images from side 1, align and export component
  • in new project import masked images from side 2, align and export component
  • create a new project and import both components
  • in 1Ds view select first component, select all its images and set Yes in Lock pose for continue
  • in 1Ds view select second component, select all its images and set Yes in Lock pose for continue
  • align images
    image
3

I will give your steps a go. The table has standard motion tracking symbols on it. Its a plastic item 2ft h x 1ft w x 1ft l. The object also has markers on it.

Thank you

4

Well I tried this on the first side…

  • import side 1 set of images, align, create model and export masks

since the table was involved, it recreated the object no problem, and i exported the masks

  • open new project, import masked images from side 1, align and export component

with the table gone, i tried to align images but it now broke up the reconstruction into 3 components that i have to use control points to piece together

this is my point, is there a way for the “alignment” data from the first step to carry over to the second step?

5

Hi, you can export also the alignment (Export/Registration), but then there won’t be used the masks.
Also, you can try export the XMPs and then use the relative registration option.
As the images are not correctly aligned with masks, then it is possible that there is not sufficient overlap between images.