With the use of opencv and its aruco module this script allows for calibration of a video camera, and distance calculation of a charuco tag given its size. The advantage of using the aruco library is it accounts for viewing angle and lens distortion. This provide a more accurate measurement and does not require the object to be viewed head on.
This is simply a script that runs the aruco.calibrateCameraCharuco function. It requires command line arguments of the cam_index as well as a filename for the output numpy matrix. The more diverse viewing angles the more accurate the results that why there is a pause included for 3 seconds after the board is detected for you to move the camera to a different angle. The script capture 15 images before running the calibration and provides a progress bar.
This loads a camera matrix created by the calibration script, and implements an instance of a camera class. This was done to allow for use of threading in future applications. the class requires you to set the variables on lines 87 - 90 however you can use the provided values for lines 89 and 90 if you print out and use test_marker0.jpg at the size it is.
name = "the name you gave the output of the calibration script"
cam_index = "camera your using"
length_constant = "since our calibration is in a unit of pixels this is a constant for pixel to feet conversion"
known_width = "this is the width of our aruco marker in pixels"The distance is calculated for each side of the square and the value returned is an average of the 4. We were able to get readings with +/- 1cm accuracy with a standard definition camera.
- Add velocity calculation
- Make the units consistent removing the need for conversion