Forum

Author Topic: Unit precision in millimetres for area and volume  (Read 1668 times)

Landy

  • Newbie
  • *
  • Posts: 36
    • View Profile
Unit precision in millimetres for area and volume
« on: November 15, 2019, 03:44:13 AM »
I have a number of models constructed in Metashape that are built from small objects (centimetre scale) photographed with a light tent and turntable setup.  I have figured out how to add a scale to the model with markers and scalebar.  The problem is that when I ask Metashape to calculate surface area and volume of the mesh, it gives the result in cubic metres and square metres respectively, and only gives precision down to 6 decimal places.  The sixth decimal place in cubic metres translates down to 1 cubic centimetre, so most of the models in question all report a volume of 1 cubic centimetre - so there is essentially no precision (see attached screen capture).  What I really need in volume in cubic millimetres, but there seems to be no option to tell Metashape to use millimetres as the local coordinate system or to provide more than six decimal places for the cubic metres result.

I suspect what I am asking for here is a feature request.  In the meantime, is there any workaround for this issue?  Does anyone know of any third party program that can calculate surface areas and volumes at millimetre precision from 3D models?  One workaround I have just thought of is to set a false scale (at a factor of 10^3) and then divide the volume result by 10^9 - just about to test this.
cheers
Bill

Landy

  • Newbie
  • *
  • Posts: 36
    • View Profile
Re: Unit precision in millimetres for area and volume
« Reply #1 on: November 15, 2019, 05:13:13 AM »
Addendum:  I have noticed that when the volume drops below 1 cm^3 (0.000001 m^3), the reported volume is then quoted in scientific notation (×10^-7) which then reveals the extra precision required.  Could I request that an option for the calculation is always reported in this fashion?  I managed to get meaningful numbers by the trick of giving a false (1000×) scale for the objects, but it would be handy not to have to do this.
cheers
Bill