ringgrid 0.5.6

Pure-Rust detector for coded ring calibration targets
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268

use nalgebra::{Matrix3, Vector3};

use super::super::DivisionModel;
use super::*;
use crate::DetectedMarker;
use crate::conic::fit_ellipse_direct;

fn circle_conic(radius: f64) -> Matrix3<f64> {
    Matrix3::new(1.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, -(radius * radius))
}

fn project_conic(q_plane: &Matrix3<f64>, h: &Matrix3<f64>) -> Matrix3<f64> {
    let h_inv = h.try_inverse().expect("invertible homography");
    h_inv.transpose() * q_plane * h_inv
}

/// Sample N points on a conic boundary.
fn sample_conic_boundary(q: &Matrix3<f64>, n: usize) -> Vec<[f64; 2]> {
    let q_sym = 0.5 * (q + q.transpose());
    let coeffs = crate::conic::ConicCoeffs([
        q_sym[(0, 0)],
        2.0 * q_sym[(0, 1)],
        q_sym[(1, 1)],
        2.0 * q_sym[(0, 2)],
        2.0 * q_sym[(1, 2)],
        q_sym[(2, 2)],
    ]);
    let e = coeffs.to_ellipse().expect("should be an ellipse");
    e.sample_points(n)
}

fn synthetic_h() -> Matrix3<f64> {
    Matrix3::new(1.12, 0.21, 321.0, -0.17, 0.94, 245.0, 8.0e-4, -6.0e-4, 1.0)
}

fn projected_center(h: &Matrix3<f64>) -> [f64; 2] {
    let p = h * Vector3::new(0.0, 0.0, 1.0);
    [p[0] / p[2], p[1] / p[2]]
}

/// Build a synthetic marker with edge points from projected circles.
///
/// When `distort` is provided, points are distorted to simulate lens distortion.
/// The edge points stored are the *distorted* image-space points (what a camera
/// would observe). The ellipse params are fitted to these distorted points.
fn make_synthetic_marker(
    h: &Matrix3<f64>,
    r_inner: f64,
    r_outer: f64,
    n_points: usize,
    distort: Option<&DivisionModel>,
) -> DetectedMarker {
    let q_inner = project_conic(&circle_conic(r_inner), h);
    let q_outer = project_conic(&circle_conic(r_outer), h);

    let mut outer_pts = sample_conic_boundary(&q_outer, n_points);
    let mut inner_pts = sample_conic_boundary(&q_inner, n_points);

    if let Some(model) = distort {
        outer_pts = outer_pts
            .iter()
            .filter_map(|p| model.distort_point(*p))
            .collect();
        inner_pts = inner_pts
            .iter()
            .filter_map(|p| model.distort_point(*p))
            .collect();
    }

    let outer_ellipse = fit_ellipse_direct(&outer_pts);
    let inner_ellipse = fit_ellipse_direct(&inner_pts);

    let center = projected_center(h);
    DetectedMarker {
        id: Some(0),
        confidence: 1.0,
        center,
        ellipse_outer: outer_ellipse,
        ellipse_inner: inner_ellipse,
        edge_points_outer: Some(outer_pts),
        edge_points_inner: Some(inner_pts),
        ..DetectedMarker::default()
    }
}

#[test]
fn division_model_identity_is_noop() {
    let model = DivisionModel::identity(640, 480);
    let p = [123.45, 67.89];
    let u = model.undistort_point(p);
    assert!(
        (u[0] - p[0]).abs() < 1e-10,
        "x: got {}, expected {}",
        u[0],
        p[0]
    );
    assert!(
        (u[1] - p[1]).abs() < 1e-10,
        "y: got {}, expected {}",
        u[1],
        p[1]
    );
}

#[test]
fn division_model_roundtrip() {
    for &lambda in &[-1e-7, -5e-8, 1e-8, 1e-7] {
        let model = DivisionModel::centered(lambda, 1280, 960);
        let p = [400.0, 300.0];
        let u = model.undistort_point(p);
        let d = model.distort_point(u).expect("distort should succeed");
        assert!(
            (d[0] - p[0]).abs() < 1e-8,
            "roundtrip failed for lambda={}: dx={}",
            lambda,
            (d[0] - p[0]).abs()
        );
        assert!(
            (d[1] - p[1]).abs() < 1e-8,
            "roundtrip failed for lambda={}: dy={}",
            lambda,
            (d[1] - p[1]).abs()
        );
    }
}

#[test]
fn objective_zero_for_perfect_circles() {
    let h = synthetic_h();
    let marker = make_synthetic_marker(&h, 4.0, 7.0, 64, None);
    let edge_data = vec![(
        marker.edge_points_outer.unwrap(),
        marker.edge_points_inner.unwrap(),
    )];
    let image_center = [320.0, 240.0];
    let obj = super::objective::conic_consistency_objective(0.0, &edge_data, image_center, 0.0);
    assert!(
        obj < 1e-6,
        "objective for perfect circles should be near zero, got {}",
        obj
    );
}

#[test]
fn objective_minimum_near_true_lambda() {
    let true_lambda = -1e-6;
    let distort_model = DivisionModel::centered(true_lambda, 1000, 1000);

    let homographies: Vec<Matrix3<f64>> = vec![
        Matrix3::new(1.12, 0.21, 100.0, -0.17, 0.94, 80.0, 8.0e-4, -6.0e-4, 1.0),
        Matrix3::new(0.95, -0.10, 850.0, 0.08, 1.05, 900.0, 5.0e-4, 3.0e-4, 1.0),
        Matrix3::new(1.05, 0.15, 900.0, -0.12, 0.90, 100.0, -3.0e-4, 7.0e-4, 1.0),
        Matrix3::new(0.88, 0.05, 120.0, 0.03, 1.10, 880.0, 6.0e-4, -2.0e-4, 1.0),
        Matrix3::new(1.20, -0.08, 750.0, 0.14, 0.85, 800.0, -5.0e-4, 4.0e-4, 1.0),
        Matrix3::new(0.92, 0.18, 180.0, -0.05, 0.98, 150.0, 4.0e-4, -5.0e-4, 1.0),
    ];

    let edge_data: Vec<super::objective::MarkerEdgeData> = homographies
        .iter()
        .map(|hi| {
            let m = make_synthetic_marker(hi, 8.0, 14.0, 128, Some(&distort_model));
            (m.edge_points_outer.unwrap(), m.edge_points_inner.unwrap())
        })
        .collect();

    let image_center = [500.0, 500.0];
    let obj_zero =
        super::objective::conic_consistency_objective(0.0, &edge_data, image_center, 0.0);
    let obj_true =
        super::objective::conic_consistency_objective(true_lambda, &edge_data, image_center, 0.0);
    let obj_wrong =
        super::objective::conic_consistency_objective(-true_lambda, &edge_data, image_center, 0.0);

    assert!(
        obj_true < obj_zero,
        "objective at true lambda ({:.3e}) should be less than at zero ({:.3e})",
        obj_true,
        obj_zero
    );
    assert!(
        obj_true < obj_wrong,
        "objective at true lambda ({:.3e}) should be less than at wrong sign ({:.3e})",
        obj_true,
        obj_wrong
    );
}

#[test]
fn golden_section_finds_quadratic_min() {
    let f = |x: f64| (x - 0.3).powi(2);
    let (x_min, f_min) = super::optimizer::golden_section_minimize(f, 0.0, 1.0, 50);
    assert!(
        (x_min - 0.3).abs() < 1e-10,
        "expected min near 0.3, got {}",
        x_min
    );
    assert!(f_min < 1e-18);
}

#[test]
fn estimate_returns_none_for_few_markers() {
    let markers: Vec<DetectedMarker> = vec![];
    let config = SelfUndistortConfig {
        enable: true,
        min_markers: 6,
        ..Default::default()
    };
    let result = estimate_self_undistort(&markers, [640, 480], &config, None);
    assert!(result.is_none());
}

#[test]
fn estimate_recovers_lambda_synthetic() {
    let true_lambda = -1e-6;
    let image_w = 1000u32;
    let image_h = 1000u32;
    let distort_model = DivisionModel::centered(true_lambda, image_w, image_h);

    let homographies: Vec<Matrix3<f64>> = vec![
        Matrix3::new(1.12, 0.21, 100.0, -0.17, 0.94, 80.0, 8.0e-4, -6.0e-4, 1.0),
        Matrix3::new(0.95, -0.10, 850.0, 0.08, 1.05, 900.0, 5.0e-4, 3.0e-4, 1.0),
        Matrix3::new(1.05, 0.15, 900.0, -0.12, 0.90, 100.0, -3.0e-4, 7.0e-4, 1.0),
        Matrix3::new(0.88, 0.05, 120.0, 0.03, 1.10, 880.0, 6.0e-4, -2.0e-4, 1.0),
        Matrix3::new(1.20, -0.08, 750.0, 0.14, 0.85, 800.0, -5.0e-4, 4.0e-4, 1.0),
        Matrix3::new(0.92, 0.18, 180.0, -0.05, 0.98, 150.0, 4.0e-4, -5.0e-4, 1.0),
        Matrix3::new(1.00, 0.12, 700.0, -0.09, 1.02, 600.0, 2.0e-4, -3.0e-4, 1.0),
        Matrix3::new(0.97, -0.15, 300.0, 0.11, 0.91, 50.0, -4.0e-4, 6.0e-4, 1.0),
    ];

    let markers: Vec<DetectedMarker> = homographies
        .iter()
        .enumerate()
        .map(|(i, hi)| {
            let mut m = make_synthetic_marker(hi, 8.0, 14.0, 128, Some(&distort_model));
            m.id = Some(i);
            m
        })
        .collect();

    let config = SelfUndistortConfig {
        enable: true,
        lambda_range: [-5e-6, 5e-6],
        max_evals: 60,
        min_markers: 6,
        improvement_threshold: 0.001,
        min_abs_improvement: 0.0,
        ..SelfUndistortConfig::default()
    };

    let result = estimate_self_undistort(&markers, [image_w, image_h], &config, None)
        .expect("should find result");

    let lambda_err = (result.model.lambda - true_lambda).abs();
    assert!(
        lambda_err < 2e-6,
        "expected lambda near {:.3e}, got {:.3e} (err={:.3e})",
        true_lambda,
        result.model.lambda,
        lambda_err
    );
    assert!(
        result.objective_at_lambda < result.objective_at_zero,
        "optimized objective ({:.3e}) should be less than baseline ({:.3e})",
        result.objective_at_lambda,
        result.objective_at_zero
    );
    assert!(result.applied, "should be marked as applied");
}