It helps in finding the extrinsic parameters of the camera which in turn helps in finding the position of the camera in a 3D pose with respect to the user defined world coordinates. Helps in setting up a global truth for robotic systems.
Installing CameraCalibrator
git clone https://github.com/jacob-02/CameraCalibrator
pip install -r requirements.txtTo deploy this project run
cd CameraCalibrator
python3 extrinsic_with_checkboard.py
python3 extrinsic_with_click.pyThe second command calibrates the camera position with respect to image points detected by checkboard corner detector function in opencv. The checkboard size, object points with respect to a world frame and if the image which we are using to calibrate is a .jpg or .png must be edited.
The third command calibrates the camera position with respect to image points selected in an image that is fed and the points are chosen by selecting on the image with a mouse pointer.
Common things to be edited in both the files are:
- Object Points with respect to a Global Origin
- Distortion Coefficient of the Camera
- Camera Matrix of the Camera
The Object Points must be measured by the user and input into the code. Distortion coeffs and Camera matrix can be found my running the following 2 codes:
cd images
python3 calibrator.py
python3 image_save.pyThe third command helps in saving multiple images of a checkboard whose size must be defined by the user correctly.
After clicking a suitable number of images, minimum 10, the user can run the second command that will help in giving the camera matrix and the distortion coefficients. This can then be inputted into the main files and give the user the translation and rotational vectors (radians and degrees).