I think that the axes in the last of tutorial2 are misleading. Usually rgb corresponds to x y z axes. That would mean that approach, axis, binormal correspond to the x,y,z axes respectively. However in the current implementation approach-> x , binormal->y and axis->z. For the image to be correct axis should point upwards instead of downwards and be coloured blue and also binormal should be painted green. I am just pointing this out because it gave me a lot of trouble to find the correct axes and had to look in the gpg code to find out. On the other hand very good job overall , I implemented this on a fetch robot and it works like a charm it can grasp almost any object on a table. Thank you !
atenpas
commented
6 years ago
Hey,
I think that the axes in the last
of tutorial2 are misleading. Usually rgb corresponds to x y z axes. That would mean that approach, axis, binormal correspond to the x,y,z axes respectively. However in the current implementation approach-> x , binormal->y and axis->z. For the image to be correct axis should point upwards instead of downwards and be coloured blue and also binormal should be painted green. I am just pointing this out because it gave me a lot of trouble to find the correct axes and had to look in the gpg code to find out. On the other hand very good job overall , I implemented this on a fetch robot and it works like a charm it can grasp almost any object on a table. Thank you !