pyfisheye.internal.utils package¶

Submodules¶

pyfisheye.internal.utils.check_shapes.check_shapes(param_names_and_rules)¶

A decorator for checking the shapes of input elements.

Parameters:: param_names_and_rules (dict[str, str]) – A dictionary containing each parameter to be checked and the corresponding rule. Only objects having ‘shape’ attributes are checked.
Returns:: The decorator.

pyfisheye.internal.utils.common.build_companion_matrix(coeffs)¶

Construct a companion matrix for a batch of polynomials. The eigenvalues of each matrix: are the roots of the corresponding polynomial.

Parameters:: coeffs (ndarray) – The coefficients in ascending order, optionally batched for multiple polynomials (of the same degree).
Return type:: ndarray

pyfisheye.internal.utils.common.compute_image_radius(image_width, image_height, distortion_centre)¶

Compute the maximum L2 norm between a pixel in the image and the distortion centre: (for normalization).

Parameters:

Return type:

float

Returns:

The image radius being the maximum L2 norm between any point on the image and the distortion centre.

pyfisheye.internal.utils.common.generate_pattern_world_coords(num_rows, num_cols, pattern_tile_size_x, pattern_tile_size_y=None)¶

Generate a num_rows*num_cols,3 array containing the pattern coordinates in world space: with Z = 0. They are returned in row-major ordering starting at (0, 0, 0).

Parameters:

num_rows (int) – The number of rows in the pattern - the number of corners in one column.
num_cols (int) – The number of columns in the pattern - the number of corners in one row.
pattern_tile_size_x (float) – The width of one tile in the pattern in metres.
pattern_tile_size_y (Optional[float]) – The height of one tile in the pattern in metres. Leave as None to default to pattern_tile_size_x.

Return type:

ndarray

Returns:

num_rows*num_cols,3 array of pattern world coordinates with Z=0. (XYZ) ordering.

pyfisheye.internal.utils.common.get_3d_transformation(extr)¶

Compute the third column of the rotation matrix and return a 3x3 rotation matrix: and 3, translation vector.

Parameters:: extr (ndarray) – The extrinsic parameters as some ndarray with 3,3 in the last two dimensions.
Return type:: tuple[ndarray, ndarray]
Returns:: The rotation matrices and translation vectors for each of the transformations, shape=N*,3,3 and N*,3

pyfisheye.internal.utils.common.get_logger()¶

pyfisheye.internal.utils.common.unnormalize_coefficients(coefficients, image_radius)¶

Convert the coefficients provided, which work on normalized values of rho.

Parameters:

coefficients (ndarray) – The 5 coefficients for the omnidirectional camera model.
image_radius (float) – The image radius in pixels.

Return type:

ndarray

Returns:

The 5 coefficients, unnormalized.