Forum

Author Topic: Point Cloud Delete Points by Confidence  (Read 3832 times)

claytonb8298

  • Newbie
  • *
  • Posts: 13
    • View Profile
Point Cloud Delete Points by Confidence
« on: March 21, 2024, 06:21:09 PM »
Hello,
When computing the point cloud in Metashape we have the program calculate confidence of each point.  You can then view the confidence of points in the model by color, hot colors have lower confidence and cool colors have higher confidence, i.e. red = is the lowest confidence and blue is the highest, and the order goes as the rainbow would, red, orange, yellow, green, blue.

You can then filter what is seen and selectable by that confidence and therefore delete the points of the lowest or lower confidence to begin cleaning up the point cloud noise.

The issue I am having is when I set the point cloud to display points of confidence between 0 - 4 and then select what is shown and delete, this delete process never deletes all those points, even when I have selected a window that obviously includes them all.  Even more frustrating, as I change my view point or view elevation, more points show up within the area I just deleted.  Meaning, I can never be certain I have deleted all the points that are within my chosen confidence interval.  This continues and continues until I get frustrated and give up.

Of course, I am hoping @Alexey Pasumansky  has the answer.  But I am truly hopeful for an answer from anyone.  :)

Thank you, Clayton

Paulo

  • Hero Member
  • *****
  • Posts: 1500
    • View Profile
Re: Point Cloud Delete Points by Confidence
« Reply #1 on: March 21, 2024, 06:54:54 PM »
Hello Clayton,

to avoid this iterative selection process, check that Visible Selection is unticked in the selection drop down button...

If it is ticked then selection would only select points that are visible from model point of view...

Hope this solves it
Best Regards,
Paul Pelletier,
Surveyor

claytonb8298

  • Newbie
  • *
  • Posts: 13
    • View Profile
Re: Point Cloud Delete Points by Confidence
« Reply #2 on: May 16, 2024, 01:23:03 AM »
Awesome! this didn't.  thank you.