Calligator

An easy-to-use library for calibrating cameras in python

The calibration library should follow these requirements:

• be able to detect points from checkerboard or charuco board easily
• save intermediate points somewhere, so that optimization can restart quickly
• optimize extrinsics and intrinsics jointly using LM method
• start with opencv intrinsics calibration if possible
• support calibration with arbitrary correspondence points using bundle adjustment
• output common diagnostics for calibration
  • reprojection error
  • length of checkerboard grids
  • picture of calibrated camera positions
  • undistorted images to check intrinsics
• triangulation functions
• end to end function to calibrate full set of cameras from list of videos

Nice to haves

• reject outlier points
• add objective in optimization to match the object points of calibration board

Functions to implement

Camera

• distort_points
• undistort_points
• project
• undistort_image
• calibrate_images
• calibrate_points
• calibrate_videos

CameraGroup

• distort_points
• undistort_points
• project
• triangulate
• reprojection_error
• calibrate_images
• calibrate_points
• calibrate_videos

CalibrationObject

• draw
• detect_image
• detect_images
• detect_video
• get_object_points
• estimate_pose