camera-calibration

To perform ROS 2 camera calibration, you capture images of a checkerboard target to estimate intrinsic parameters and distortion coefficients. The process outputs a camera_info.yaml file for direct integration into your ROS 2 camera drivers.

You need camera recalibration whenever the lens, mounting, or focus changes. Without updated intrinsic parameters and distortion coefficients, 3D reconstruction, map alignment, and localization accuracy will suffer during ROS 2 perception tasks.

Yes, it supports both monocular and stereo camera calibration in ROS 2. Monocular calibration estimates intrinsic parameters, while stereo calibration calculates the baseline and rectification matrices required for depth perception and stereo vision.

The calibration process generates a camera_info.yaml file containing intrinsic parameters, distortion coefficients, and stereo rectification matrices. These artifacts allow ROS 2 camera drivers to consume the data and produce accurately rectified image outputs.

When selecting a checkerboard for camera calibration, you must define the board size and square dimensions accurately. Proper selection, along with adequate image coverage and lighting, ensures accurate intrinsic parameter estimation and successful rectification validation.