Crate calib_targets_chessboard 

Plain chessboard detector built on top of calib-targets-core.

Quickstart

use calib_targets_chessboard::{ChessboardDetector, ChessboardParams};
use calib_targets_core::Corner;

let params = ChessboardParams::default();
let detector = ChessboardDetector::new(params);

let corners: Vec<Corner> = Vec::new();
let result = detector.detect_from_corners(&corners);
println!("detected: {}", result.is_some());

New algorithm (graph-based, perspective-aware):

1. Filter strong ChESS corners.
2. Estimate two approximate global grid axes (u, v) from ChESS orientations.
3. Estimate a base spacing from nearest-neighbor distances.
4. For each corner, find up to 4 neighbors (right/left/up/down) based on:
   • distance ~ base spacing,
   • direction roughly along ±u or ±v (orientation-based),
5. Build an undirected 4-connected grid graph from these neighbor relations.
6. BFS each connected component, assign integer coordinates (i, j).
7. For each component, compute width×height and completeness.
8. Keep the best component that matches expected_rows/cols (up to swap) and completeness threshold.

Structs

ChessboardDebug

ChessboardDetectionResult

ChessboardDetector

Simple chessboard detector using ChESS orientations + grid fitting in (u, v) space.

ChessboardParams

Parameters specific to the chessboard detector.

GridGraphDebug

GridGraphNeighborDebug

GridGraphNodeDebug

GridGraphParams

RectifiedBoardView

RectifiedMeshView

Enums

MeshWarpError

RectifyError

Functions

rectify_from_chessboard_result

rectify_mesh_from_grid

Build a rectified “board view” by piecewise homographies per grid cell. This is robust to lens distortion because it does not assume a single global H.