Rigid3d Tutorial Guide

p_a = np.array([0, 1, 0]) p_b = T[:3,:3] @ p_a + T[:3,3] print(p_b) # [0., 0., 0.] If you have ( T_bc ) and ( T_ab ), the transform from ( a ) to ( c ) is:

[ p_B = R_AB \cdot p_A + t_AB ]

[ T_ac = T_ab \cdot T_bc ]

SE3d T_ba = T_ab.inverse();

In robotics, computer vision, and 3D graphics, the ability to represent rotations and translations in 3D space is fundamental. The Rigid3D object (often found in libraries like Sophus , Eigen , geometry_msgs , or tf2 ) is the industry-standard way to do this. Unlike a 4x4 homogeneous matrix, Rigid3D separates rotation (SO(3)) and translation, offering better numerical stability and mathematical clarity. rigid3d tutorial

Rigid3D typically stores these as a unit quaternion (for rotation) and a 3-vector (for translation). For this tutorial, we'll use the Sophus::SE3d (C++) or scipy.spatial.transform.Rotation (Python). If you're working with ROS, tf2::Transform is analogous. C++ (Eigen + Sophus) #include <sophus/se3.hpp> #include <Eigen/Core> #include <iostream> using Sophus::SE3d; using Eigen::Vector3d; using Eigen::Quaterniond; Python (NumPy + SciPy or transforms3d) import numpy as np from scipy.spatial.transform import Rotation as R # Or use `from transforms3d.quaternions import quat2mat` 3. Creating a Rigid3D Object Let’s create a transformation that represents: rotate 90° about Z-axis, then translate by (1, 0, 0). p_a = np