MATLAB Program for Camera Intrinsic Parameter Calibration Using Checkerboard Pattern

This MATLAB program performs camera intrinsic parameter calibration using checkerboard patterns and implements Zhang's classic calibration algorithm. The implementation involves several key steps: first, users need to prepare a checkerboard pattern and capture multiple images of it from different angles using the camera. The program then automatically detects checkerboard corner points in the images using built-in MATLAB functions like detectCheckerboardPoints(), which employs Harris corner detection and sub-pixel refinement for accurate corner localization. The calibration algorithm proceeds by establishing correspondences between 3D world points (from the checkerboard pattern) and 2D image points. It computes the homography matrix for each image view and then solves for initial camera parameters using linear least-squares estimation. The implementation includes non-linear optimization (typically using Levenberg-Marquardt algorithm) to refine parameters and minimize reprojection errors. Key computed intrinsic parameters include focal length (fx, fy), principal point coordinates (cx, cy), and lens distortion coefficients (radial and tangential). The program generates calibration results with accuracy metrics and visualization of reprojection errors. This calibration provides essential foundation for subsequent computer vision tasks such as 3D reconstruction, stereo vision, and image measurement applications. The implementation follows MATLAB's Camera Calibrator app workflow but provides programmable access to all calibration steps, allowing customization and batch processing capabilities for advanced users.