Closed lema18 closed 5 years ago
@lema18 PnP uses 3D world points and their projections on the camera to compute the camera pose wrt to the world. So here R and t is the relation between world frame and camera frame.
My comment on line 262 is wrong. Sorry about that.
thanks
In vo.cpp at line 302 at poseEstimationPnP function.. why do you use curr_->T_wc = transRt2T(R, t).inv(); instead of curr_->T_wc =ref_->T_w_c * transRt2T(R, t).inv(); ?