Tasks #978

NaN when converting in singularities

Added by Anonymous about 12 years ago. Updated over 10 years ago.

Status:NewStart date:04/03/2012
Priority:UrgentDue date:
Assignee:-% Done:

20%

Category:-Estimated time:1.00 hour
Target version:rci0.5

Description

Example matrix that results in NaNs in compact axis-angle conversion (in z-poles):

homogen matrix: 
 [-0.995,    -0.102,    0.0114,    -0.507;
 -0.102,    0.971,    -0.218,    -0.512;
 0.0111,    -0.218,    -0.976,    0.19;

Associated revisions

Revision 566
Added by anordman about 12 years ago

Quick fix for #978, perturbing the rotation matrix a tiny bit if in a pole, to avoid 'nan's.
refs #978

History

#1 Updated by Anonymous about 12 years ago

here another example with higher precision:
[-0.9996185303, -0.02710843459, 0.005286611617, -0.4500545561;
-0.02711678855, 0.9269508123, -0.3742015958, -0.5852395296;
0.005243589636, -0.3742021918, -0.9273323417, 0.1688518077;

#2 Updated by Anonymous about 12 years ago

and another one:

[-0.9984720349, -0.0552579537, 0.0004264877061, -0.5783057809;
-0.0552588217, 0.9983872175, -0.01301411726, -0.005522137508;
0.0002933336364, -0.01301779971, -0.9999152422, 0.1988631487;

#3 Updated by Anonymous about 12 years ago

  • % Done changed from 0 to 20

Quick fixed, but not fixed in general. Need to understand what happens inside the poles and how to handle this (in each of the conversions).
Damn you rotations!

#4 Updated by Anonymous over 10 years ago

  • Target version set to rci0.4

#5 Updated by Anonymous over 10 years ago

  • Target version changed from rci0.4 to rci0.5

Also available in: Atom PDF