Projecting CPs (control points) in given image

Completed

Follow

Iconem

• August 02, 2017 17:21

A nice addition with Control Points (CPs) placement would be
1- (optional) to be able to place them in the sarse point cloud directly for a broad positioning.
2- once a CP is already placed in some images, and a user wants to place it in a new image, he should be able to let RC "make a guess", ie project this component CP in that camera. Then, the user can adjust the positioning of the CP more accurately. This would be useful to set CPs among a single component, but wont make easier the positioning of CPs across more than one component.

[edit] in new RC update, projection of pivot point might be handy for this feature

1

• 

  Götz Echtenacher

  • August 13, 2017 16:40

  +10!
  I've also wished for that more than once!
  Even better would be to add GCPs in a georeferenced model in the estimated location...
  
  How do you suggest show pivot could be used for that?
  I haven't had the time to try this feature out yet...

  0

  Comment actions Permalink
• 

  Götz Echtenacher

  • August 14, 2017 07:59

  I just got a chance to try out Pivot Point.
  It displays the center point (pivot) of the 3D view (set by double clicking on the point cloud) in a 2D image.
  Super practical if you try to find a feature in difficult terrain.
  It will save LOADS of time!!!
  Well done, RC-Team!

  0

  Comment actions Permalink
• 

  Götz Echtenacher

  • August 18, 2017 11:50

  Just discovered something by chance.
  Once you pull any old GCP (2D) into an empty one, it shows up immediately in 3D.
  So that means at least that functionality - displaying GCPs in the 3D window - is already there in theory.
  
  So please: Display all imported GCPs in 3D right away!
  
  Edit: Displayed grey when empty, blue when ok and red when there's a problem...

  0

  Comment actions Permalink
• 

  Jennifer Cross

  • September 26, 2017 08:14

  Even more - how about a function to project existing CPs into all photos where the point cloud indicates they should be viewable (with the px error indicated when the point is manually repositioned). 

  This would save a bunch of "should I put a tag that" in the workflow and most of the tagging process just is visual confirmation that all the CPs in the image are sitting where you expect.

  You could use the same function to re-project the CP based on alignment data (force all error 0px) and then we could shift the CP in any badly aligned images (reset the errors then align again and see where the adjustments are large)

  The last few updates have made adding CP much easier!  Thank you

  0

  Comment actions Permalink

Please sign in to leave a comment.