README_calibration.md

Calibration of CapeReviso Camera System

Prerequisites

Instrinsic calibration with chessboard markers

Some routines to capture these on a system (including IDS-Cameras) are included with the tkDNN-Fork, so calibration can be done with the same sttings as image capturing later on.

Distort the background image of the location using the intrinsic calibration

python undistort_single_image.py -i calibration/calibration02_i -f vhs-calibration/visagx2_background_image_2022-11-30_11-41-11.jpg

Convert a csv-File to opencv-file containing obj-points

python utmcoordinatetransform.py vhs.cvs will result in a YAML-file with object points

Extrinsic calibration

python ImageCoordinatesTool2.py -p vhs-calibration/visagx2_background_image_2022-11-30_11-41-11_undistorted.jpg will result in a YAML-file with image points If the same locations are chosen in the same order as in the previous step converting the csv file, the resulting files can work together.

Edit config

image_pts_path = vhs-calibration/visagx2_background_image_2022-11-30_11-41-11_undistorted_2022_12_12_1516.yml object_pts_path = vhs-calibration/vhs_obj_points_2022-12-13_1315.yml calibration_image_path = vhs-calibration/visagx2_background_image_2022-11-30_11-41-11_undistorted.jpg

Check calibration by running

python test_global_utm_calib.py -p vhs-calibration/config_vhs_wg_noshow.ini