Python Projective Camera Model

Features

• camera intrinsic and extrinsic parameters handling
• various lens distortion models
• model persistence
• projection of camera coordinates to an image
• conversion of image coordinates on a plane to camera coordinates
• visibility handling

Installation

pip install git+https://github.com/smidm/camera.py.git

Usage

Setup camera and project point [0, 0, 0]:

import camera
import numpy as np
import math
c = camera.Camera()
c.set_K_elements(1225.0, math.pi / 2, 1, 480, 384)
R = np.array(
    [[-0.9316877145365, -0.3608289515885, 0.002545329627547],
     [-0.1725273110187, 0.4247524018287, -0.8888909933995],
     [0.3296724908378, -0.8263880720441, -0.4579894432589]])
c.set_R(R)
c.set_t(np.array([[-1.365061486465], [3.431608806127], [17.74182159488]]))

print c.world_to_image(np.array([[0., 0., 0.]]).T)

Load camera parameters from a file:

import camera
import numpy as np
c = camera.Camera()
c.load('camera_01.yaml')
c.world_to_image(np.array([[0., 0., 0.]]).T)

Documentation

http://python-projective-camera-model.readthedocs.io

Development

To test the package run:

$ ./run_tests.sh

The module is missing some (essential) functionality (search for TODO comments). I appreciate pull requests that fill the gaps.

License

MIT