`scipy.spatial.transform._rigid_transform:RigidTransform.from_matrix`

source: /scipy/spatial/transform/_rigid_transform.py:478

Signature

staticmethoddef from_matrix(matrix: ArrayLike) → RigidTransform

Summary

Initialize from a 4x4 transformation matrix.

Parameters

matrix : array_like, shape (..., 4, 4): Transformation matrices. Each matrix[..., :, :] represents a 4x4 transformation matrix.

Returns

transform : `RigidTransform` instance: Object containing the rigid transform represented by the input matrix.

Notes

4x4 rigid transformation matrices are of the form

[       tx]
[   R   ty]
[       tz]
[ 0 0 0  1]

where R is a 3x3 rotation matrix and t = [tx, ty, tz] is a 3x1 translation vector. As rotation matrices must be proper orthogonal, the rotation component is orthonormalized using singular value decomposition before initialization.

Examples

from scipy.spatial.transform import RigidTransform as Tf
import numpy as np

✓

Creating a transform from a single matrix:

m = np.array([[0, 1, 0, 2],
              [0, 0, 1, 3],
              [1, 0, 0, 4],
              [0, 0, 0, 1]])
tf = Tf.from_matrix(m)
tf.as_matrix()
tf.single

✓

True

Creating a transform from an N-dimensional array of matrices:

m = np.tile(np.eye(4), (2, 5, 1, 1))  # Shape (2, 5, 4, 4)
tf = Tf.from_matrix(m)
tf.shape
tf.single
len(tf)

✓

Matrices with a rotation component that is not proper orthogonal are orthogonalized using singular value decomposition before initialization:

tf = Tf.from_matrix(np.diag([2, 2, 2, 1]))
tf.as_matrix()

✓

array([[1., 0., 0., 0.],
       [0., 1., 0., 0.],
       [0., 0., 1., 0.],
       [0., 0., 0., 1.]])

Aliases

scipy.spatial.transform.RigidTransform.from_matrix