Pipeline Overview

At a high level, the workflow looks like this:

1. Input corners: supply a list of calib_targets_core::Corner values, typically from a ChESS detector.
2. Estimate grid axes: cluster corner orientations to get two dominant grid directions.
3. Build a grid graph: connect corners that plausibly lie on the same grid lines.
4. Assign integer coordinates: BFS the graph to produce (i, j) grid indices.
5. Select the best board: choose the best connected component that matches expected size.
6. Rectify (optional): compute a global homography or mesh warp to build a rectified view.
7. Decode markers (optional): decode per-cell directly, or scan a rectified grid if you need a full rectified image.
8. Align board (optional): map markers to a known layout and assign corner IDs.

Output types are standardized in calib-targets-core as TargetDetection with LabeledCorner values. Higher-level crates enrich that output with additional metadata (inliers, marker detections, rectified views).