This is an unofficial implementation of Z. Zhang, IEEE Transactions 2000. "A flexible new technique for camera calibration"
All the data I used are in directory: data/left
You can run prob1_6.py to see the camera calibration results and the undistrot images results
You can run main.py to see the camera calibration implementation without using function cv2.calibrateCamera().
pic_data.py: Prepare data.
dlt.py: Caculate Homography of each images.
intrinsics.py: Caculate the intrinsics matrix.
extrinsics.py: Caculate the extrinsics matrix.
-
The intrinsics caculate by using OpenCV API :
A =
(534.07, 0, 0)
(0, 534.12, 0)
(341.53, 232.95, 1) -
The intrinsics caculate by using my implementation:
A =
(532.806, 0, 0)
(1.521, 529.066, 0)
(350.359, 225.710, 1) -
Note by Chenyu Gao, Northwestern Polytechnical University, Xi'an, Apr, 2018
-
Any questions: chenyugao.cs@gmail.com