#30 singular matrix error can be caused by several points with same coordinate forming a plane

[Andrew Hudak]

If the x-coordinates (or y-coordinates) are the same, it causes the interpolated surface to be a vertical plane, which produces an infinite solution for the curvature coefficient and causes a crash. Currently MCC-LIDAR only checks for this condition if 3 points produce the plane, but this case makes it apparent that the vertical plane test needs to be applied to a greater number of points. Or perhaps there's another more elegant/efficient solution to this problem?

[Andrew Hudak]

• attachment _singular_matrix_error_screenshot.jpg_ added

[Jimm Domingo]

It's hard to read the coordinate data from the screenshot. Could you copy and paste the text in the error message into a new comment for this ticket? Thanks.

[Jimm Domingo]

• owner set to jimm_domingo
• status changed from new to assigned

A. Hudak emailed me a Word document with the screenshot in it, so I could read the error message. Here it is:

Error: singular matrix while computing thin plate spline
  Spline control points:
    (524341.72, 5189006.49, 867.63)
    (524341.72, 5189004.93, 867.71)
    (524341.72, 5189006.25, 867.65)
    (524341.72, 5189005.5, 867.65)
    (524341.72, 5189004.32, 867.77)
    (524341.72, 5189006.71, 867.62)
    (524341.72, 5189005.22, 867.68)
    (524341.72, 5189006.93, 867.58)
    (524341.72, 5189004.62, 867.76)
    (524341.72, 5189007.12, 867.57)
    (524341.72, 5189006.01, 867.65)
    (524341.72, 5189005.76, 867.66)

[Jimm Domingo]

• status changed from assigned to closed
• resolution set to fixed

Fixed in revision r167.

[Jimm Domingo]

• milestone set to 1.0 (rc2)