I had a bug in the analytic jacobian that I don't really understand, but maybe someone knows how this is happening.
For the analytic jacobian, at some point the coordinates must be transformed back to world space.
There are multiple possibilities..
When I transform the P3D points back to world space and do my math I got a weird skew on the jacobian matrix.
When I transform the V3D vectors back to world orientation and do my math I got much cleaner results.
I would expect them to be the same..
All that comes to mind is that it might be a numerical issue, but this does not seem right. But I don't have any other idea what could have happen there..
In the end it might also be a bug, which is awfully close to the expected values.
Hi
I had a bug in the analytic jacobian that I don't really understand, but maybe someone knows how this is happening.
For the analytic jacobian, at some point the coordinates must be transformed back to world space. There are multiple possibilities.. When I transform the P3D points back to world space and do my math I got a weird skew on the jacobian matrix. When I transform the V3D vectors back to world orientation and do my math I got much cleaner results. I would expect them to be the same..
All that comes to mind is that it might be a numerical issue, but this does not seem right. But I don't have any other idea what could have happen there.. In the end it might also be a bug, which is awfully close to the expected values.