Skip to main content
P
PE_98950691-Visitor
1-Visitor
April 7, 2021
Question

2D images to 3D surface point cloud

  • April 7, 2021
  • 2 replies
  • 2144 views

Has any one used MathCad to take 2D images (e.g., bitmaps) and render a 3D point cloud from them? Is this do-able with MathCad? 

2 replies

F
Fred_Kohlhepp23-Emerald I
23-Emerald I
April 7, 2021

"Old" mathcad could read image files into matrices, these commands are still in Prime.  Reading two or more 2D images and discerning point-by-point coordinates from them would be theoretically feasible, but mathematically challenging.  I don't recall seeing anything so ambitious! 

P
PE_98950691-VisitorAuthor
1-Visitor
April 7, 2021

Well...I'm thinking of a constrained situation where the 2D images are all a priori correlated to each other...

F
Fred_Kohlhepp23-Emerald I
23-Emerald I
April 8, 2021

I'm thinking of a constrained situation where the 2D images are all a priori correlated to each other...

 

Clearly there would have to be a correlation between the images.  In drafting you need three orthogonal views, front, end, and top.  (XZ view, YZ view, and XY view.)  Clear datums and scale markers between views would be required to develop three dimensional coordinates (X, Y, Z) from those three views for each point of interest.  So you get two coordinates for a point in each view and (hopefully) the same value shows up between views; 

  • Z is the same between XZ and YZ,
  • X is the same between XZ and XY,
  • Y is the same between XY and YZ

Challenging!

P
PP_99035901-Visitor
1-Visitor
April 15, 2021

The overall process flow of the system from 2D Image into 3D point cloud. To test on the proposed methodology, two real objects are captured as images for trial run in order to obtain results. The first is an image of a hot cup cover and the second is an image of component from grass trimmer.

Terms & ConditionsAccessibility statement