camera.camera — CV Camera Functions

The camera.camera module provides methods for working with idealized cameras and look vector intersections.

New in version 0.1.0.

autocnet.camera.camera.camera_from_f(F)[source]

Estimate a camera matrix using a fundamental matrix.

Parameters

f (ndarray) – (3,3) fundamental matrix or autocnet Fundamental Matrix object

Returns

p1 – Estimated camera matrix

Return type

ndarray

autocnet.camera.camera.compute_epipoles(f)[source]

Compute the epipole and epipolar prime

Parameters

f (ndarray) – (3,3) fundamental matrix or autocnet Fundamental Matrix object

Returns

  • e (ndarray) – (3,1) epipole

  • e1 (ndarray) – (3,3) epipolar prime matrix

autocnet.camera.camera.idealized_camera()[source]

Create an idealized camera transformation matrix

Returns

(3,4) with diagonal 1

Return type

ndarray

autocnet.camera.camera.projection_error(p1, p, pt, pt1)[source]

Based on Hartley and Zisserman p.285 this function triangulates image correspondences and computes the reprojection error by back-projecting the points into the image.

This is the classic cost function (minimization problem) into the gold standard method for fundamental matrix estimation.

Parameters
  • p1 (ndarray) – (3,4) camera matrix

  • p (ndarray) – (3,4) idealized camera matrix in the form np.eye(3,4)

  • pt (dataframe or ndarray) – of homogeneous coordinates in the form (x_{i}, y_{i}, 1)

  • pt1 (dataframe or ndarray) – of homogeneous coordinates in the form (x_{i}, y_{i}, 1)

Returns

reproj_error – (n, 1) vector of reprojection errors

Return type

ndarray

autocnet.camera.camera.triangulate(pt, pt1, p, p1)[source]

Given two sets of homogeneous coordinates and two camera matrices, triangulate the 3D coordinates. The image correspondences are assumed to be implicitly ordered.

References

[Hartley2003]

Parameters
  • pt (ndarray) – (n, 3) array of homogeneous correspondences

  • pt1 (ndarray) – (n, 3) array of homogeneous correspondences

  • p (ndarray) – (3, 4) camera matrix

  • p1 (ndarray) – (3, 4) camera matrix

Returns

coords – (4, n) projection matrix

Return type

ndarray