typing
module-attribute
Type alias for 2D vector inputs.
Accepts
- Vector2 instances
- NumPy arrays with shape (2,)
- Lists: [x, y]
- Tuples: (x, y)
- Any sequence with 2 float elements
module-attribute
Vector3Like = Union[
Vector3,
ArrayLike,
Sequence[float],
tuple[float, float, float],
list[float],
]
Type alias for 3D vector inputs.
Accepts
- Vector3 instances
- NumPy arrays with shape (3,)
- Lists: [x, y, z]
- Tuples: (x, y, z)
- Any sequence with 3 float elements
module-attribute
Type alias for 6D vector inputs.
Accepts
- Vector6 instances
- NumPy arrays with shape (6,)
- Lists: [x1, x2, x3, x4, x5, x6]
- Tuples: (x1, x2, x3, x4, x5, x6)
- Any sequence with 6 float elements
module-attribute
QuaternionLike = Union[
Quaternion,
ArrayLike,
Sequence[float],
tuple[float, float, float, float],
list[float],
]
Type alias for quaternion inputs.
Accepts
- Quaternion instances
- NumPy arrays with shape (4,) - [w, x, y, z]
- Lists: [w, x, y, z]
- Tuples: (w, x, y, z)
- Any sequence with 4 float elements representing quaternion components
module-attribute
Type alias for general rotation inputs.
Accepts
- All QuaternionLike inputs
- 3x3 rotation matrices as NumPy arrays
- Any array-like object representing rotations
module-attribute
Type alias for Region of Interest (ROI) inputs.
Accepts
- Roi instances
- NumPy arrays with shape (4,)
- Lists: [xmin, ymin, xmax, ymax]
- Tuples: (xmin, ymin, xmax, ymax)
- Any sequence with 4 float elements representing bounding box coordinates
module-attribute
Type alias for single point inputs (2D or 3D).
module-attribute
PointsLike = Union[
ArrayLike,
Sequence[Sequence[float]],
list[list[float]],
list[tuple[float, ...]],
]
Type alias for multiple points inputs.
Accepts
- NumPy arrays with shape (N, 2) or (N, 3)
- Lists of lists: [[x1, y1], [x2, y2], ...]
- Lists of tuples: [(x1, y1), (x2, y2), ...]
- Any nested sequence representing multiple points
module-attribute
Type alias for 3x3 matrix inputs.
Accepts
- NumPy arrays with shape (3, 3)
- Nested lists: [[a, b, c], [d, e, f], [g, h, i]]
- Any nested sequence representing a 3x3 matrix
module-attribute
Type alias for 4x4 matrix inputs.
Accepts
- NumPy arrays with shape (4, 4)
- Nested lists: 4x4 structure
- Any nested sequence representing a 4x4 matrix
module-attribute
Type alias for trajectory inputs.
Accepts
- NumPy arrays with shape (M, T, D) where:
- M = number of modes
- T = number of timesteps
- D = spatial dimensions (usually 3)
- Triple-nested sequences with same structure
module-attribute
CameraIntrinsicLike = Union[
CameraIntrinsic,
ArrayLike,
Sequence[Sequence[float]],
list[list[float]],
]
Type alias for camera parameter inputs.
Accepts
- CameraIntrinsic instances
- NumPy arrays with shape (3, 3)
- Nested sequences of float values
module-attribute
Type alias for camera distortion inputs.
Accepts
- CameraDistortion instances
- NumPy arrays with shape (5,)
- Sequence of float values
module-attribute
Type alias for scalar numeric inputs.
Accepts
- Python int or float
- NumPy scalar types
Bases: ndarray
A 3x3 camera intrinsic matrix with validation.
This class ensures that the input array is a 3x3 matrix and raises a ValueError if it is not. It can be constructed from a 3x3 array or 9 elements array.
Note that for non-camera, the array can be empty.
Examples:
>>> i = CameraIntrinsic(np.eye(3)) # OK
>>> i = CameraIntrinsic([1, 0, 0, 0, 1, 0, 0, 0, 1]) # OK
>>> i = CameraIntrinsic([[1, 0, 0], [0, 1, 0], [0, 0, 1]]) # OK
>>> i = CameraIntrinsic([]) # OK
>>> i = CameraIntrinsic(np.eye(2)) # ValueError
>>> i = CameraIntrinsic([1, 0, 0, 0, 1, 0, 0, 0]) # ValueError
Bases: ndarray
A 1D array representing camera lens distortion coefficients.
This class represents the distortion parameters for a camera lens following the OpenCV distortion model. It supports arrays of length 4, 5, 8, 12, or 14 elements, corresponding to different distortion models.
The distortion coefficients follow the OpenCV convention: (k1, k2, p1, p2[, k3[, k4, k5, k6[, s1, s2, s3, s4[, τx, τy]]]])
Coefficient meanings
- k1, k2, k3, k4, k5, k6: Radial distortion coefficients
- p1, p2: Tangential distortion coefficients
- s1, s2, s3, s4: Thin prism distortion coefficients
- τx, τy: Tilted sensor distortion coefficients
Supported array lengths
- 4 elements: (k1, k2, p1, p2) - Basic radial and tangential
- 5 elements: (k1, k2, p1, p2, k3) - Extended radial distortion
- 8 elements: (k1, k2, p1, p2, k3, k4, k5, k6) - Rational model
- 12 elements: (k1, k2, p1, p2, k3, k4, k5, k6, s1, s2, s3, s4) - With thin prism
- 14 elements: (k1, k2, p1, p2, k3, k4, k5, k6, s1, s2, s3, s4, τx, τy) - Full model
Note that for non-camera, the array can be empty.
Examples:
>>> d = CameraDistortion([0, 0, 0, 0]) # 4 elements: basic model
>>> d = CameraDistortion([0, 0, 0, 0, 0]) # 5 elements: with k3
>>> d = CameraDistortion([1, 2, 3, 4, 5]) # 5 elements: valid coefficients
>>> d = CameraDistortion([1, 2, 3, 4, 5, 6, 7, 8]) # 8 elements: rational model
>>> d = CameraDistortion([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12]) # 12 elements: with thin prism
>>> d = CameraDistortion([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]) # 14 elements: full model
>>> d = CameraDistortion([]) # Empty array for non-camera
>>> d = CameraDistortion([1, 2, 3]) # ValueError: invalid length
>>> d = CameraDistortion([1, 2, 3, 4, 5, 6]) # ValueError: invalid length
Bases: ndarray
A 3x3 matrix with validation.
Examples:
>>> matrix = Matrix3x3([[1, 2, 3], [4, 5, 6], [7, 8, 9]]) # OK
>>> matrix = Matrix3x3(np.eye(3)) # OK
>>> matrix = Matrix3x3([[1, 2], [4, 5], [7, 8]]) # ValueError
>>> matrix = Matrix3x3(np.eye(2)) # ValueError
Bases: ndarray
A 4x4 matrix with validation.
Examples:
>>> matrix = Matrix4x4([[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16]]) # OK
>>> matrix = Matrix4x4(np.eye(4)) # OK
>>> matrix = Matrix4x4([[1, 2], [4, 5], [7, 8]]) # ValueError
>>> matrix = Matrix4x4(np.eye(2)) # ValueError
Bases: Quaternion
A quaternion class that wraps the PyQuaternion class.
This wrapper exists to provide a consistent and explicit quaternion representation.
Examples:
>>> q = Quaternion(1, 2, 3, 4)
>>> q
Quaternion(1.000000, 2.000000, 3.000000, 4.000000)
>>> q.conjugate()
Quaternion(1.000000, -2.000000, -3.000000, -4.000000)
>>> q.norm()
5.477226
>>> q.inverse()
Quaternion(0.181818, -0.363636, -0.545455, -0.727273)
>>> q * q.inverse()
Quaternion(1.000000, 0.000000, 0.000000, 0.000000)
Bases: tuple
A 4-element tuple representing a region of interest (ROI).
This class ensures that the array always has the correct shape and value order. It can be constructed from any array-like object that can be converted to a tuple of length 4, or from individual numeric arguments.
Examples:
>>> roi = Roi(10, 20, 30, 40) # OK
>>> roi = Roi((10, 20, 30, 40)) # OK
>>> roi = Roi([10, 20, 30, 40]) # OK
>>> roi = Roi([10, 20]) # ValueError: ROI must be 4-elements
>>> roi = Roi([40, 30, 20, 10]) # ValueError: ROI must be xmin <= xmax && ymin <= ymax
property
Return the xy offset from the image origin at the top left corner.
Bases: BaseVector
A 2-element numpy array with validation.
This class ensures that the array always has exactly 2 elements. It can be constructed from any array-like object that can be converted to a 2-element numpy array.
Examples:
>>> v = Vector2([1, 2]) # OK
>>> v = Vector2(np.array([1, 2])) # OK
>>> v = Vector2(1, 2) # OK
>>> v = Vector2([1, 2, 3]) # ValueError
Bases: BaseVector
A 3-element numpy array with validation.
This class ensures that the array always has exactly 3 elements. It can be constructed from any array-like object that can be converted to a 3-element numpy array.
Examples:
>>> v = Vector3([1, 2, 3]) # OK
>>> v = Vector3(np.array([1, 2, 3])) # OK
>>> v = Vector3(1, 2, 3) # OK
>>> v = Vector3([1, 2]) # ValueError
Bases: BaseVector
A 6-element numpy array with validation.
This class ensures that the array always has exactly 6 elements. It can be constructed from any array-like object that can be converted to a 6-element numpy array.
Examples:
>>> v = Vector6([1, 2, 3, 4, 5, 6]) # OK
>>> v = Vector6(np.array([1, 2, 3, 4, 5, 6])) # OK
>>> v = Vector6(1, 2, 3, 4, 5, 6) # OK
>>> v = Vector6([1, 2]) # ValueError