Crate calib_targets_core 

Core types and utilities for calibration target detection.

This crate is intentionally small and purely geometric. It does not depend on any concrete corner detector or image type.

Quickstart

use calib_targets_core::{Corner, TargetDetection, TargetKind};
use nalgebra::Point2;

let corner = Corner {
    position: Point2::new(10.0, 20.0),
    orientation: 0.0,
    orientation_cluster: None,
    strength: 1.0,
};

let detection = TargetDetection {
    kind: TargetKind::Chessboard,
    corners: Vec::new(),
};

println!("{:?} {}", corner.position, detection.corners.len());

Includes

• Homography estimation and warping helpers.
• Lightweight grayscale image views and sampling.
• Grid alignment and target detection types.

Structs

Corner

Canonical 2D corner used by all target detectors.

GrayImage

GrayImageView

GridAlignment

A grid alignment dst = transform(src) + translation.

GridCoords

Integer grid coordinates (i, j) in board space.

GridTransform

Integer 2D grid transform (a 2×2 matrix) used for aligning detected grids to a board model.

Homography

LabeledCorner

A corner that is part of a detected target, with optional ID info.

OrientationClusteringParams

Parameters for orientation clustering.

OrientationClusteringResult

Result of orientation clustering.

OrientationHistogram

RectifiedView

TargetDetection

One detected target (board instance) in an image.

Enums

RectToImgMapper

TargetKind

The kind of target that a detection corresponds to.

Constants

GRID_TRANSFORMS_D4

The 8 dihedral transforms D4 on the integer grid.

Functions

cluster_orientations

Cluster ChESS orientations into two dominant directions on [0, π).

compute_orientation_histogram

Compute smoothed orientation histogram for debug/visualization.

estimate_grid_axes_from_orientations

Estimate a dominant grid axis from orientations using a double-angle mean.

estimate_homography_rect_to_img

Estimate H such that: p_img ~ H * p_rect

homography_from_4pt

Compute H such that: dst ~ H * src (projective), using 4 point correspondences.

init_tracing

init_with_level

Install the simple logger with the provided level filter.

sample_bilinear

sample_bilinear_fast

sample_bilinear_u8

warp_perspective_gray

Warp into rectified image: for each dst pixel, map to src via H_img_from_rect and sample.