OpenCV Python based camera calibration

Features

Corner Detection

Undistortion Images

Calculate Reprojection Errors

Show histogram Errors

Calibration Results

{
    "reprojection_error": 1.2273468645903203,
    "camera_matrix": [
        [
            1422.117605057742,
            0.0,
            924.5633003737133
        ],
        [
            0.0,
            1423.204374674482,
            667.6183041323897
        ],
        [
            0.0,
            0.0,
            1.0
        ]
    ],
    "dist_coeffs": [
        4.612616555464993,
        152.56623164609113,
        0.039943271525320366,
        -0.008947777240770066,
        -228.0106478053562,
        4.592880469213394,
        151.6208067211176,
        -221.65838356166944,
        0.0010495594571102234,
        0.011289450598695816,
        -0.038397785485157175,
        -0.03986920764202555
    ]
}

Environment

conda create -n calib python=3.11

pip install -r requirements.txt

How to use

1. Print Checkerboard images (assets/checkerboard.png) to A4

2. Recording and then move video file to assets/calibration_video.mp4

3. Extract video frames

   python extract_video.py

4. Run calibration

   python camera_calibration.py # simple run
   
   python camera_calibration_high.py # Best quality