ringgrid (Python)

Python bindings for the ringgrid detector (PyO3 + maturin).

Install

From PyPI:

pip install ringgrid

With plotting helpers:

pip install "ringgrid[viz]"

From source (repository checkout):

pip install maturin
maturin develop -m crates/ringgrid-py/Cargo.toml --release

Fast Start: Generate `target_spec.json` + Printable SVG/PNG/DXF

Installed-package target generation is available directly from import ringgrid via the typed TargetLayout API:

from pathlib import Path
import ringgrid

target = ringgrid.TargetLayout.coded_hex(
    pitch_mm=8.0,
    rows=15,
    long_row_cols=14,
    outer_radius_mm=4.8,
    inner_radius_mm=3.2,
    ring_width_mm=1.152,
)

Path("target_spec.json").write_text(target.to_spec_json())
target.write_svg(Path("target_print.svg"), margin_mm=5.0)
target.write_png(Path("target_print.png"), dpi=600.0, margin_mm=5.0)
target.write_dxf(Path("target_print.dxf"))

Key knobs:

API	What it controls	Typical value
`TargetLayout.coded_hex(...)` / `.coded_rect(...)`	Lattice geometry (`pitch_mm`, dims, radii, ring width)	`8.0`, `15`, `14`, `4.8`, `3.2`, `1.152`
`TargetLayout.plain_hex(...)` / `.plain_rect(...)`	Lattice geometry plus `dots=` (auto origin fiducials)	`14.0`, `24`, `24`, `5.6`, `2.8`, `dots=True`
`TargetLayout.default_hex()` / `.rect_24x24()` / `TargetLayout(...)`	Frozen presets and the full compositional constructor	—
`write_svg(..., margin_mm=...)`	Extra white border around the printable page	`3-10`
`write_png(..., dpi=...)`	PNG raster resolution and embedded print metadata	`300` or `600`
`write_png(..., include_scale_bar=...)`	Include or omit the default scale bar	`True`

Outputs:

target_spec.json (schema v6)
target_print.svg
target_print.png
target_print.dxf (2D CAD, millimeters — for laser/CNC fabrication)

Equivalent paths for the same geometry (identical SVG/PNG/DXF; all write a v6 target_spec.json):

Rust CLI: ringgrid gen <recipe.toml> --out <dir> — a small TOML/JSON target recipe (start from ringgrid example --name hex_coded) renders the same target_spec.json + SVG/PNG/DXF
Python script from the repo: tools/gen_target.py for the same geometry
Rust API: TargetLayout::coded_hex(...) / TargetLayout::new(...) plus write_json_file, write_target_svg, write_target_png, and write_target_dxf

Legacy v4 board_spec.json files still load — ringgrid.TargetLayout.from_json(...) auto-migrates v4 and v5 to v6. (See migration notes.)

Load this target in Python:

from pathlib import Path
import ringgrid

target = ringgrid.TargetLayout.from_json(Path("tools/out/target_faststart/target_spec.json"))
cfg = ringgrid.DetectConfig(target)
detector = ringgrid.Detector(cfg)
# Convenience default (config-free): detector = ringgrid.Detector.from_target(target)

If you are working from a repository checkout and also need synthetic images or ground truth, the repo tools under tools/ still provide the combined generation/evaluation workflow. The installed package target-generation API is for target JSON + printable SVG/PNG/DXF only. The repo-level tools/gen_target.py is a thin wrapper over this same installed-package surface.

Complete target-generation tutorial and full flag reference:

https://vitalyvorobyev.github.io/ringgrid/book/target-generation.html

Target layouts

TargetLayout is the typed, first-class target model — the Python mirror of the Rust ringgrid.target.v6 schema. It expresses hex and rectangular lattices, coded and plain (uncoded) markers, and optional origin fiducials.

import ringgrid

# Presets (geometry comes from the native library — no duplicated constants):
hex_target = ringgrid.TargetLayout.default_hex()        # 15-row coded hex, 203 markers
rect_target = ringgrid.TargetLayout.rect_24x24()   # 24x24 plain rect + origin dots

# One constructor per target-matrix row, all taking plain scalars:
coded_hex = ringgrid.TargetLayout.coded_hex(
    pitch_mm=8.0, rows=15, long_row_cols=14,
    outer_radius_mm=4.8, inner_radius_mm=3.2, ring_width_mm=1.152,
)
coded_rect = ringgrid.TargetLayout.coded_rect(14.0, 20, 20, 4.8, 3.2, 1.152)

# Plain targets take `dots=`: True auto-places a rotation-asymmetric origin-dot
# triad in the lattice gaps (absolute board frame); False omits it (the board is
# then labeled up to lattice symmetry and must be detected completely).
rect_plain = ringgrid.TargetLayout.plain_rect(14.0, 24, 24, 5.6, 2.8, dots=True)
hex_plain = ringgrid.TargetLayout.plain_hex(8.0, 15, 14, 4.8, 3.2, dots=False)

# Dot placement (lattice gaps + rotational-symmetry validation) is computed
# natively — you never hand-author dot coordinates. The full compositional
# constructor mirrors the v6 schema verbatim if you need it:
explicit = ringgrid.TargetLayout(
    name="my_rect",
    lattice=ringgrid.RectGeometry(rows=24, cols=24, pitch_mm=14.0),
    marker=ringgrid.RingGeometry(outer_radius_mm=5.6, inner_radius_mm=2.8),
    coding=ringgrid.Plain(),
)

# Detector / DetectConfig accept a TargetLayout directly:
detector = ringgrid.Detector.from_target(hex_target)
# ...or: ringgrid.Detector(ringgrid.DetectConfig(hex_target))

to_dict() / from_dict() round-trip the v6 schema verbatim; from_json(...) loads v6 (or legacy v5 / v4, auto-migrated) text or a file path; to_spec_json() returns canonical, validated v6 JSON. Invalid geometry raises ValueError with the native error message:

lattice: ringgrid.LatticeGeometry = ringgrid.HexGeometry(15, 14, 8.0)
coding: ringgrid.MarkerCoding = ringgrid.Coded16(ring_width_mm=1.152)

restored = ringgrid.TargetLayout.from_json("target.json")
assert ringgrid.TargetLayout.from_dict(restored.to_dict()) == restored

Features

Typed TargetLayout (ringgrid.target.v6) — hex/rect lattices, coded/plain markers, origin fiducials
Native TargetLayout target generation for canonical spec JSON + printable SVG/PNG/DXF
Native Detector API with NumPy input support
Slim DetectionResult model objects with JSON round-trips
Opt-in DetectionDiagnostics channel (detector.detect_with_diagnostics(...)) for per-marker fit/decode internals, edge points, and homography RANSAC stats
Optional plotting helpers in ringgrid.viz (pip install ringgrid[viz])

Input Rules

Detector.detect(...) accepts:
- np.ndarray with dtype=uint8 and shape (H, W) (grayscale)
- np.ndarray with dtype=uint8 and shape (H, W, 3|4) (RGB/RGBA, auto-converted to grayscale)
- image file path (str or pathlib.Path)
Other dtypes/shapes raise TypeError.

Proposal-Only Diagnostics

You can run just the proposal stage and inspect the heatmap used for proposal localization:

import ringgrid
from ringgrid import viz

proposals = ringgrid.propose("photo.png")
diagnostics = ringgrid.propose_with_heatmap("photo.png")

print(len(proposals))
print(diagnostics.heatmap.shape)  # (H, W), float32

viz.plot_proposal_diagnostics(
    image="photo.png",
    diagnostics=diagnostics,
    out="proposal_diagnostics.png",
)

If you want proposal generation to follow the detector's existing scale tuning, use the detector-bound methods instead:

target = ringgrid.TargetLayout.default_hex()
cfg = ringgrid.DetectConfig(target)
detector = ringgrid.Detector(cfg)

diagnostics = detector.propose_with_heatmap("photo.png")

ProposalResult.heatmap is the post-Gaussian-smoothed vote accumulator that the proposal stage uses for thresholding and NMS.

Full tutorial and repo tool workflow:

Book: Proposal Diagnostics
python tools/plot_proposal.py --image ... --gt ... --out ...

DetectConfig Field Guide

DetectConfig is the full Python tuning surface for Detector.detect(...), detect_adaptive(...), and detect_multiscale(...).

import ringgrid

target = ringgrid.TargetLayout.default_hex()
cfg = ringgrid.DetectConfig(target)

# Section properties return copies: mutate, then reassign.
decode = cfg.decode
decode.codebook_profile = "extended"
decode.min_decode_margin = 2
cfg.decode = decode

# Or use convenience aliases for common one-field tweaks.
cfg.completion_enable = False
cfg.decode_min_confidence = 0.4

snapshot = cfg.to_dict()
# Stage tuning nests under "advanced" in the wire view.
print(snapshot["advanced"]["decode"]["codebook_profile"])  # "extended"

How Python DetectConfig behaves:

cfg.target is the constructor input and stays read-only. It is not included in cfg.to_dict().
cfg.to_dict() returns the resolved native wire view. That is the easiest way to inspect the exact config the Rust detector will use. Its top-level keys are marker_scale, circle_refinement, self_undistort, and advanced; every stage-tuning section (decode, inner_fit, outer_fit, completion, proposal, id_correction, …) lives under the advanced object.
Section getters such as cfg.decode, cfg.inner_fit, and cfg.self_undistort stay flat on the Python DetectConfig object — they return copies. Reassign the section after editing it, or use a convenience alias such as cfg.decode_min_margin = 2. (The flat Python accessors map onto the nested advanced wire fields automatically.)
cfg.marker_scale defaults to 14-66 px outer diameter and re-derives the scale-coupled search windows when you replace it.
Target geometry derives cfg.marker_spec.r_inner_expected and cfg.decode.code_band_ratio. For TargetLayout.default_hex(), those resolve to 0.48809522 and 0.74404764.
cfg.circle_refinement uses the Python enum ringgrid.CircleRefinementMethod, while cfg.to_dict()["circle_refinement"] stores the native wire strings "ProjectiveCenter" or "None".

Default marker_scale derivations for DetectConfig(TargetLayout.default_hex()) (stage sections shown as nested advanced wire keys):

advanced.proposal.r_min — spacing-aware, max(0.15 * spacing_min_px, 2.0) -> 3.0310888
advanced.proposal.r_max — spacing-aware, min(0.45 * spacing_max_px, 1.35 * radius_max_px) -> 42.86825
advanced.proposal.min_distance — derived from marker spacing and diameter prior
advanced.edge_sample.r_max = 2.0 * radius_max_px -> 66.0
advanced.outer_estimation.search_halfwidth_px = max(max((radius_max_px - radius_min_px) * 0.5, 2.0), 13.0) -> 13.0
advanced.completion.roi_radius_px = clamp(0.75 * nominal_diameter_px, 24.0, 80.0) -> 30.0
advanced.projective_center.max_correction_shift_px — stays None ("auto"): the gate falls back to the nominal marker diameter (40.0 px here) at use time

Deeper theory and Rust-side derivation details:

Surface Map

Surface	Type	What it controls
`cfg.target`	`TargetLayout`	Target geometry used to derive geometry-coupled defaults
`cfg.marker_scale`	`MarkerScalePrior`	Expected marker diameter range in working pixels
`cfg.proposal`	`ProposalConfig`	Scharr-vote proposal generation
`cfg.edge_sample`	`EdgeSampleConfig`	Radial edge sampling limits and density
`cfg.outer_estimation`	`OuterEstimationConfig`	Outer-radius hypothesis generation from radial peaks
`cfg.marker_spec`	`MarkerSpecConfig`	Board-driven ring geometry assumptions
`cfg.outer_fit`	`OuterFitConfig`	Outer ellipse fit acceptance and scoring
`cfg.inner_fit`	`InnerFitConfig`	Inner ellipse fit acceptance and penalties
`cfg.decode`	`DecodeConfig`	Code-band sampling and decode strictness
`cfg.seed_proposals`	`SeedProposalConfig`	Seed-injected proposals for multi-pass flows
`cfg.projective_center`	`ProjectiveCenterConfig`	Projective-center recovery gates
`cfg.completion`	`CompletionConfig`	Homography-guided recovery of missing IDs
`cfg.ransac_homography`	`RansacConfig`	Global homography fitting thresholds
`cfg.self_undistort`	`SelfUndistortConfig`	Division-model self-undistort estimation
`cfg.id_correction`	`IdCorrectionConfig`	Hex-lattice ID verification and recovery
`cfg.inner_as_outer_recovery`	`InnerAsOuterRecoveryConfig`	Recovery when outer fit locked onto the inner ring

Top-Level Controls

Property	Default	Practical notes
`cfg.target`	constructor input	Read-only target layout. Replace the whole config if you need a different target.
`cfg.circle_refinement`	`ringgrid.CircleRefinementMethod.PROJECTIVE_CENTER`	Use `NONE` for raw ellipse centers, or keep `PROJECTIVE_CENTER` for the accuracy-oriented default.
`cfg.dedup_radius`	`6.0`	Final marker merge radius in pixels. Raise only if duplicate fits survive; lower if nearby valid markers merge incorrectly.
`cfg.max_aspect_ratio`	`3.0`	Rejects very elongated ellipses. Tighten when false positives are obviously non-circular; loosen only for extreme perspective.
`cfg.use_global_filter`	`True`	Enables homography-based outlier rejection. Turn it off when debugging local fits or when you want to inspect raw pre-homography detections.

Convenience Aliases

Alias	Expands to	When to use it
`cfg.completion_enable`	`cfg.completion.enable`	Quick toggle for homography-guided completion
`cfg.self_undistort_enable`	`cfg.self_undistort.enable`	Quick toggle for division-model self-undistort inside `detect()`
`cfg.inner_fit_required`	`cfg.inner_fit.require_inner_fit`	Promote missing inner fits from soft penalty to hard reject
`cfg.homography_inlier_threshold_px`	`cfg.ransac_homography.inlier_threshold`	Tighten or loosen global homography inlier gating
`cfg.decode_min_margin`	`cfg.decode.min_decode_margin`	Reject ambiguous decodes more aggressively
`cfg.decode_max_dist`	`cfg.decode.max_decode_dist`	Limit how many bit errors a decode may contain
`cfg.decode_min_confidence`	`cfg.decode.min_decode_confidence`	Raise or lower overall decode strictness without rebuilding the section

`marker_scale`

This is the highest-leverage tuning section. Replacing cfg.marker_scale recomputes the scale-coupled defaults in proposal, edge_sample, outer_estimation.search_halfwidth_px, and completion.roi_radius_px; the projective_center.max_correction_shift_px "auto" fallback also tracks it.

Field	Default	Practical notes
`diameter_min_px`	`14.0`	Minimum expected outer diameter in working pixels. Raise it if markers are never tiny.
`diameter_max_px`	`66.0`	Maximum expected outer diameter. Narrow it to cut false positives; widen it only if markers truly get larger.

If markers span a very wide range, prefer detect_adaptive(...) or detect_multiscale(...) over one very wide marker_scale.

`proposal`

Controls the first Scharr-gradient voting stage that proposes candidate marker centers before local fitting.

Field	Default	Practical notes
`r_min`	derived -> `3.0310888`	Minimum vote radius. Lower only for genuinely tiny markers. Re-derived (spacing-aware) from `cfg.marker_scale` and board geometry.
`r_max`	derived -> `42.86825`	Maximum vote radius. Raise only if markers exceed your current size prior. Re-derived (spacing-aware) from `cfg.marker_scale` and board geometry.
`grad_threshold`	`0.05`	Fraction of max gradient magnitude used to keep votes. Raise it in noisy scenes; lower it for low-contrast imagery.
`min_distance`	derived	Minimum distance between proposals (px). Re-derived from `cfg.marker_scale`.
`min_vote_frac`	`0.1`	Minimum accumulator peak fraction relative to the best proposal. Raise to be stricter, lower to keep weaker peaks.
`max_candidates`	`None`	Optional hard cap on proposals. Use only when you must bound runtime in cluttered scenes.
`radius_step`	`1`	Stride between voting radii. `1` (default) tests every integer radius; `2`+ subsamples (≈ halves proposal cost at `2`) but lowers recall on blurry/real scenes. The max radius is always included.

`edge_sample`

Controls the radial rays used to collect inner and outer edge evidence around a proposal.

Field	Default	Practical notes
`n_rays`	`48`	Angular sampling density. More rays improve stability on oblique markers at extra cost.
`r_max`	derived -> `66.0`	Maximum sampling radius. Re-derived from `cfg.marker_scale`.
`r_min`	`1.5`	Minimum sampling radius. Rarely changed directly.
`r_step`	`0.5`	Radial step in pixels. Lower values sample more densely but cost more.
`min_ring_depth`	`0.08`	Minimum signed edge depth kept during sampling. Raise for noisy false edges; lower for low-contrast targets.
`min_rays_with_ring`	`16`	Minimum rays that must see a valid ring-like response. Raise for stricter geometry, lower for partial occlusion.

`outer_estimation`

Controls the radial profile stage that predicts an outer radius before the ellipse fit. This stage is cheaper than a full fit, so it is a good place to reject weak hypotheses early.

Field	Default	Practical notes
`search_halfwidth_px`	derived -> `13.0`	Radius search window around the prior. Re-derived from `cfg.marker_scale` but never below the base default.
`radial_samples`	`64`	Samples per ray used to estimate radial peaks. Higher values help blurry targets at extra runtime.
`aggregator`	`"median"`	Cross-ray aggregation policy. The shipped value is the only one used in first-party docs/tests.
`grad_polarity`	`"dark_to_light"`	Expected outer-edge gradient direction. Match this only if you intentionally invert target contrast assumptions.
`min_theta_coverage`	`0.6`	Minimum fraction of rays with valid evidence. Raise it to reject partial arcs sooner.
`min_theta_consistency`	`0.35`	Minimum agreement fraction around the selected radius. Raise for stricter peak consensus.
`allow_two_hypotheses`	`True`	Lets the stage carry a strong secondary radius hypothesis into later scoring. Helpful for ambiguous profiles.
`second_peak_min_rel`	`0.85`	Relative strength required for that second hypothesis. Raise it to keep only nearly-tied alternatives.
`refine_halfwidth_px`	`1.0`	Per-ray refinement window around the selected peak. Raise slightly for blurrier edges.

`marker_spec`

Describes the expected ring geometry. The target layout drives r_inner_expected, while the remaining fields control radial/theta sampling and coverage checks.

Field	Default	Practical notes
`r_inner_expected`	derived -> `0.48809522`	Expected inner/outer radius ratio after board-geometry padding. Usually change the board geometry, not this field.
`inner_search_halfwidth`	`0.08`	Half-width around `r_inner_expected` used to search for the inner ring. Widen only if the target design itself differs.
`inner_grad_polarity`	`"light_to_dark"`	Expected inner-edge polarity. Match this only if you intentionally invert the target rendering.
`radial_samples`	`64`	Samples per theta spoke. Raise for softer gradients; lower only if you are aggressively trading accuracy for runtime.
`theta_samples`	`96`	Angular samples around the ring. More samples help oblique or partially occluded markers.
`aggregator`	`"median"`	Cross-theta aggregation policy. The shipped value is the canonical path.
`min_theta_coverage`	`0.6`	Minimum angular coverage for a valid marker profile. Lower it only for partial visibility.
`min_theta_consistency`	`0.25`	Minimum angular consistency around the chosen profile. Raise to reject uneven or contaminated profiles earlier.

`outer_fit`

Controls the final outer-ellipse fit. Nested ransac is a ringgrid.RansacFitConfig.

Field	Default	Practical notes
`min_direct_fit_points`	`6`	Minimum points required for a direct algebraic fit. Rarely tuned.
`min_ransac_points`	`8`	Minimum points before RANSAC is attempted. Lower only for severe occlusion experiments.
`ransac.max_iters`	`200`	More iterations help with heavy outliers but cost runtime.
`ransac.inlier_threshold`	`1.5`	Sampson-distance threshold in pixels. Tighten for cleaner data; loosen for blur/distortion.
`ransac.min_inliers`	`6`	Minimum inlier count accepted by the outer fit.
`ransac.seed`	`42`	Deterministic RNG seed for the fit.
`size_score_weight`	`0.15`	Weight of size agreement in the outer-hypothesis score. Raise if size priors are highly reliable.
`max_angular_gap_rad`	`1.5707963267948966`	Largest allowed missing arc gap (pi/2 by default). Lower for stricter completeness, raise for partial arcs.

`inner_fit`

Controls the inner-ellipse fit that refines geometry and influences final confidence. Nested ransac is a ringgrid.RansacFitConfig.

Field	Default	Practical notes
`min_points`	`20`	Minimum points before an inner fit is attempted. Lower only when many markers are partially cropped.
`min_inlier_ratio`	`0.5`	Required RANSAC inlier fraction. Raise for cleaner scenes, lower for blur/heavy distortion.
`max_rms_residual`	`1.0`	Maximum RMS Sampson residual. Tighten to reject sloppy inner fits.
`max_center_shift_px`	`12.0`	Largest allowed shift between outer and inner centers. Raise only if strong perspective or distortion genuinely moves the fitted center more.
`max_ratio_abs_error`	`0.15`	Maximum deviation from the expected inner/outer ratio. Tighten for well-calibrated, fixed targets.
`local_peak_halfwidth_idx`	`3`	Radial index half-width around the predicted inner peak.
`ransac.max_iters`	`200`	More iterations help when inner edges are noisy.
`ransac.inlier_threshold`	`1.5`	Inner-fit Sampson threshold in pixels.
`ransac.min_inliers`	`8`	Minimum inliers required for the inner fit.
`ransac.seed`	`43`	Deterministic RNG seed for the inner fit.
`miss_confidence_factor`	`0.7`	Confidence multiplier when the inner fit is missing. Lower values punish missing inner rings more strongly.
`max_angular_gap_rad`	`1.5707963267948966`	Largest missing inner-edge arc accepted before rejection.
`require_inner_fit`	`False`	Soft by default. Set to `True` when you want to reject any marker that lacks a trustworthy inner ellipse.

`decode`

Controls code-band sampling and codebook matching.

Field	Default	Practical notes
`codebook_profile`	`"base"`	String selector: `"base"` keeps the stable shipped IDs `0..892`; `"extended"` opts into the additive larger profile.
`code_band_ratio`	derived -> `0.74404764`	Sampling radius inside the outer ellipse. Derived from board geometry and usually not tuned directly.
`samples_per_sector`	`5`	Angular intensity samples per bit sector. Raise for blur, lower only for aggressive speed tradeoffs.
`n_radial_rings`	`3`	Radial samples across the code band. More rings improve robustness on soft edges.
`max_decode_dist`	`3`	Maximum Hamming distance accepted. Lower it to reject noisy decodes more aggressively.
`min_decode_confidence`	`0.3`	Overall decode-confidence floor. Raise this first when you want stricter decoding.
`min_decode_margin`	`1`	Rejects ambiguous ties by default. Raising it is a strong way to prefer only very clear decodes.
`min_decode_contrast`	`0.03`	Minimum sampled code-band contrast before decoding. Lower only for low-contrast images.
`threshold_max_iters`	`10`	Iteration cap for the internal 2-means threshold refinement.
`threshold_convergence_eps`	`0.0001`	Convergence epsilon for that refinement loop.

Typical decode tuning:

Too many ambiguous IDs: raise cfg.decode_min_margin or lower cfg.decode_max_dist.
Good geometry but noisy contrast: lower min_decode_contrast slightly before widening geometry gates.
Need IDs beyond the stable shipped set: set codebook_profile to "extended" explicitly.

`seed_proposals`

Controls proposal injection from already-known seed centers during multi-pass or guided workflows.

Field	Default	Practical notes
`merge_radius_px`	`3.0`	Seed/proposal merge distance. Raise only if seed centers are systematically off by several pixels.
`seed_score`	`1000000000000.0`	Score assigned to injected seeds so they survive proposal ranking.
`max_seeds`	`512`	Optional cap on consumed seeds. Use it to bound runtime when external seed lists get large.

`projective_center`

Controls the center-refinement stage used when cfg.circle_refinement == ringgrid.CircleRefinementMethod.PROJECTIVE_CENTER.

Field	Default	Practical notes
`use_expected_ratio`	`True`	Uses the board-driven inner/outer ratio as a prior in the selector. Usually leave this on.
`ratio_penalty_weight`	`1.0`	Strength of that ratio prior. Lower it if you need the selector to trust raw conic evidence more.
`max_correction_shift_px`	`None` (auto)	Maximum accepted correction jump (pre-0.8 name: `max_center_shift_px`). `None` falls back to the nominal marker diameter; explicit values are honored as-is.
`max_selected_residual`	`0.25`	Rejects unstable projective-center candidates. Raise only if valid markers are failing this gate.
`min_eig_separation`	`1e-06`	Guards against unstable conic-pencil eigenpairs. Lower only if you have evidence the default is too strict.

`completion`

Completion tries to recover missing IDs at homography-projected board locations. It only runs when a valid homography is available.

Field	Default	Practical notes
`enable`	`True`	Set to `False` to inspect only directly fit-decoded markers.
`roi_radius_px`	derived -> `30.0`	Radius of the completion search ROI. Re-derived from `cfg.marker_scale`.
`reproj_gate_px`	`3.0`	Maximum allowed distance between the fitted center and the projected board position.
`min_fit_confidence`	`0.45`	Minimum confidence for a recovered completion marker.
`min_arc_coverage`	`0.35`	Minimum fraction of rays that found both edges. Lower only for heavy occlusion.
`max_attempts`	`None`	Optional cap on attempted missing IDs.
`image_margin_px`	`10.0`	Skip projected centers too close to the image boundary.
`require_perfect_decode`	`False`	Strong safety gate for distortion-heavy scenes without a trusted mapper.
`max_radii_std_ratio`	`0.35`	Rejects fits with highly inconsistent outer radii across rays.

`ransac_homography`

Controls global homography fitting from decoded markers.

Field	Default	Practical notes
`max_iters`	`2000`	Iteration budget for the global RANSAC loop.
`inlier_threshold`	`5.0`	Pixel reprojection threshold for inliers. Tighten for cleaner data; loosen for more distortion or weak initial geometry.
`min_inliers`	`6`	Minimum correspondences accepted for a homography.
`seed`	`0`	Deterministic RNG seed for repeatable fitting.

`self_undistort`

Controls the optional one-parameter division-model self-undistort flow. cfg.self_undistort_enable = True affects Detector.detect(...), but Detector.detect_with_mapper(...) always uses the mapper you pass in instead.

Field	Default	Practical notes
`enable`	`False`	Turn on only when you want `detect()` to estimate distortion first.
`lambda_range`	`[-8e-07, 8e-07]`	Search interval for the division-model parameter. Widen only if you know distortion is stronger.
`max_evals`	`40`	Maximum objective evaluations during optimization.
`min_markers`	`6`	Minimum markers with usable edge data before self-undistort is attempted.
`improvement_threshold`	`0.01`	Relative improvement required before the estimate is considered useful.
`min_abs_improvement`	`0.0001`	Absolute improvement floor. Prevents tiny numerical wins from activating the model.
`trim_fraction`	`0.1`	Fraction of worst residuals trimmed in robust scoring.
`min_lambda_abs`	`5e-09`	Rejects near-zero solutions that do not meaningfully change the model.
`reject_range_edge`	`True`	Rejects solutions that land too close to the search-interval edge.
`range_edge_margin_frac`	`0.02`	Edge margin used by that range-edge rejection.
`validation_min_markers`	`24`	Marker count required for the homography validation pass.
`validation_abs_improvement_px`	`0.05`	Absolute reprojection improvement required by validation.
`validation_rel_improvement`	`0.03`	Relative reprojection improvement required by validation.

`id_correction`

Runs after local fit/decode to verify or recover IDs from the board's hex lattice structure.

Field	Default	Practical notes
`enable`	`True`	Leave on unless you are explicitly debugging raw decoder output.
`auto_search_radius_outer_muls`	`[2.4, 2.9, 3.5, 4.2, 5.0]`	Staged neighborhood radii for local search. Tighten only if incorrect neighbors dominate.
`consistency_outer_mul`	`3.2`	Neighborhood radius for structural consistency checks.
`consistency_min_neighbors`	`1`	Minimum neighbors required before a consistency check runs.
`consistency_min_support_edges`	`1`	Minimum supporting board-neighbor edges required to keep an ID.
`consistency_max_contradiction_frac`	`0.5`	Maximum allowed contradiction fraction before clearing an ID.
`soft_lock_exact_decode`	`True`	Protects exact decodes unless structure strongly contradicts them.
`min_votes`	`2`	Votes required to change an already-assigned ID.
`min_votes_recover`	`1`	Votes required to recover a missing ID.
`min_vote_weight_frac`	`0.55`	Minimum weighted-vote share for the winning candidate.
`h_reproj_gate_px`	`30.0`	Loose reprojection gate used by the fallback homography assignment.
`homography_fallback_enable`	`True`	Enables fallback ID recovery from a rough homography when local evidence is weak.
`homography_min_trusted`	`24`	Minimum trusted markers before that fallback is attempted.
`homography_min_inliers`	`12`	Minimum inliers required for the fallback homography.
`max_iters`	`5`	Maximum iterative correction passes.
`remove_unverified`	`False`	Default keeps the detection but clears the ID. Set to `True` to drop unverifiable markers entirely.
`seed_min_decode_confidence`	`0.7`	Minimum decode confidence used when bootstrapping trusted seeds without a homography.

`inner_as_outer_recovery`

Post-processing stage that tries to fix markers whose outer fit locked onto the inner ring.

Field	Default	Practical notes
`enable`	`True`	Leave on for the default blur-tolerant behavior.
`ratio_threshold`	`0.75`	Neighbor-radius ratio below which a marker is considered suspicious.
`k_neighbors`	`6`	Number of nearest neighbors used to estimate the expected outer radius.
`min_theta_consistency`	`0.18`	Lower-than-normal consistency gate used by this recovery path.
`min_theta_coverage`	`0.4`	Minimum angular coverage required during the re-fit.
`min_ring_depth`	`0.02`	Relaxed edge-depth gate used for blurry outer edges.
`refine_halfwidth_px`	`2.5`	Wider local radius refinement window for the recovery re-fit.
`size_gate_tolerance`	`0.25`	Prevents the relaxed recovery fit from re-locking onto the inner ring.

Typical tuning sequence:

Known scale: tighten cfg.marker_scale before touching low-level proposal thresholds.
Too many weak IDs: raise cfg.decode_min_confidence or cfg.decode_min_margin.
Good local fits but unstable global cleanup: tighten cfg.homography_inlier_threshold_px.
Distorted scenes without calibration: try cfg.self_undistort_enable = True.
Want direct detections only: set cfg.completion_enable = False.

Adaptive Detection

Use adaptive detection when marker diameter varies substantially across the image (near/far perspective, zoom changes, mixed target scales).

Which Method Should I Use?

Situation	Recommended call	Why
You do not know marker size in advance	`detector.detect_adaptive(image)`	Probes scale and auto-selects tiers
You know approximate marker diameter (px)	`detector.detect_adaptive(image, nominal_diameter_px=d)`	Skips probe and uses focused two-tier bracket around `d`
You need fixed/reproducible tier policy	`detector.detect_multiscale(image, tiers)`	Full explicit control over tiers
Marker size range is tight and runtime is priority	`detector.detect(image)`	Single-pass (fastest)

Canonical adaptive entry point is:

Detector.detect_adaptive(image, nominal_diameter_px: float | None = None)

Compatibility alias (deprecated, still supported):

Detector.detect_adaptive_with_hint(image, nominal_diameter_px=...)

Tier objects:

ScaleTier(diameter_min_px, diameter_max_px)
ScaleTiers([...])
Presets: ScaleTiers.four_tier_wide(), ScaleTiers.two_tier_standard()
Single-pass equivalent: ScaleTiers.single(MarkerScalePrior(...))

Practical Recipes

Unknown scene scale:

from pathlib import Path
import ringgrid

target = ringgrid.TargetLayout.default_hex()
detector = ringgrid.Detector.from_target(target)
image = Path("testdata/target_3_split_00.png")

result = detector.detect_adaptive(image)

Known nominal diameter (for example, ~32 px):

result = detector.detect_adaptive(image, nominal_diameter_px=32.0)

Inspect tiers used by adaptive logic (debug/repro):

tiers = detector.adaptive_tiers(image, nominal_diameter_px=32.0)
for tier in tiers.tiers:
    print(tier.diameter_min_px, tier.diameter_max_px)

# Re-run exactly those tiers
result = detector.detect_multiscale(image, tiers)

Examples

Run from repository root:

python crates/ringgrid-py/examples/basic_detect.py \
  --image testdata/target_3_split_00.png \
  --out testdata/target_3_split_00_det_py.json

python crates/ringgrid-py/examples/detect_with_camera.py \
  --image testdata/target_3_split_00.png \
  --out testdata/target_3_split_00_det_cam_py.json

python crates/ringgrid-py/examples/detect_adaptive.py \
  --image testdata/target_3_split_00.png \
  --out testdata/target_3_split_00_det_adaptive_py.json

python crates/ringgrid-py/examples/detect_multiscale.py \
  --image testdata/target_3_split_00.png \
  --tiers four_tier_wide \
  --out testdata/target_3_split_00_det_multiscale_py.json

Plotting example:

python crates/ringgrid-py/examples/plot_detection.py \
  --image testdata/target_3_split_00.png \
  --out testdata/target_3_split_00_overlay_py.png

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

The dropdown lists show the available interpreters, ABIs, and platforms. Enable javascript to be able to filter the list of wheel files.

ringgrid-0.11.0-cp314-cp314-win_amd64.whl (2.0 MB view details)

Uploaded Jul 23, 2026 CPython 3.14Windows x86-64

ringgrid-0.11.0-cp314-cp314-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.4 MB view details)

Uploaded Jul 23, 2026 CPython 3.14manylinux: glibc 2.17+ x86-64

ringgrid-0.11.0-cp314-cp314-macosx_11_0_arm64.whl (2.1 MB view details)

Uploaded Jul 23, 2026 CPython 3.14macOS 11.0+ ARM64

ringgrid-0.11.0-cp313-cp313-win_amd64.whl (2.0 MB view details)

Uploaded Jul 23, 2026 CPython 3.13Windows x86-64

ringgrid-0.11.0-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.4 MB view details)

Uploaded Jul 23, 2026 CPython 3.13manylinux: glibc 2.17+ x86-64

ringgrid-0.11.0-cp313-cp313-macosx_11_0_arm64.whl (2.1 MB view details)

Uploaded Jul 23, 2026 CPython 3.13macOS 11.0+ ARM64

ringgrid-0.11.0-cp312-cp312-win_amd64.whl (2.0 MB view details)

Uploaded Jul 23, 2026 CPython 3.12Windows x86-64

ringgrid-0.11.0-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.4 MB view details)

Uploaded Jul 23, 2026 CPython 3.12manylinux: glibc 2.17+ x86-64

ringgrid-0.11.0-cp312-cp312-macosx_11_0_arm64.whl (2.1 MB view details)

Uploaded Jul 23, 2026 CPython 3.12macOS 11.0+ ARM64

ringgrid-0.11.0-cp311-cp311-win_amd64.whl (2.0 MB view details)

Uploaded Jul 23, 2026 CPython 3.11Windows x86-64

ringgrid-0.11.0-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.4 MB view details)

Uploaded Jul 23, 2026 CPython 3.11manylinux: glibc 2.17+ x86-64

ringgrid-0.11.0-cp311-cp311-macosx_11_0_arm64.whl (2.1 MB view details)

Uploaded Jul 23, 2026 CPython 3.11macOS 11.0+ ARM64

ringgrid-0.11.0-cp310-cp310-win_amd64.whl (2.0 MB view details)

Uploaded Jul 23, 2026 CPython 3.10Windows x86-64

ringgrid-0.11.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.4 MB view details)

Uploaded Jul 23, 2026 CPython 3.10manylinux: glibc 2.17+ x86-64

ringgrid-0.11.0-cp310-cp310-macosx_11_0_arm64.whl (2.1 MB view details)

Uploaded Jul 23, 2026 CPython 3.10macOS 11.0+ ARM64

File details

Details for the file ringgrid-0.11.0-cp314-cp314-win_amd64.whl.

File metadata

Download URL: ringgrid-0.11.0-cp314-cp314-win_amd64.whl
Upload date: Jul 23, 2026
Size: 2.0 MB
Tags: CPython 3.14, Windows x86-64
Uploaded using Trusted Publishing? Yes
Uploaded via: twine/6.1.0 CPython/3.13.14

File hashes

Hashes for ringgrid-0.11.0-cp314-cp314-win_amd64.whl
Algorithm	Hash digest
SHA256	`d0a275e54a7f96e87da3879d340175d19888ea04541361d7dce876ded78c9bb4`
MD5	`d3c7b9af699c74daa146c5b79f1c8ec2`
BLAKE2b-256	`f47727293bff9799ad87abb55372633f55400721a186dcfa0bfde20d74a2016d`

Algorithm	Hash digest
SHA256	`80377fdfb7d405b2915d7ae3284bfe50672e91e4afa2d1074d26e5a6543cf0fd`
MD5	`222d404a4c0ccd172e2d0c1535f13def`
BLAKE2b-256	`8f4cc5e40e5ecd49955d38ec8eb9f9641243a1e9acec6afd18f886d32266030b`

Algorithm	Hash digest
SHA256	`80cfb058ef6478363821a2a2cd4c53f5b6ac67906ad6b8f6bbf163d7fdecf35e`
MD5	`16128905f9888dad42c504fe828aa2c3`
BLAKE2b-256	`1f7f382a9bc315cdf0890718a0e7a19aa2acf69ea8b49b6208228ef853b24789`

Algorithm	Hash digest
SHA256	`1a63d09ee8ba5da6ac228ea08a885d2a4fdddf2b332f5680548b59e32f561b30`
MD5	`e96c8c4968c27d5bbe13449fb04cd3d9`
BLAKE2b-256	`c7dae3f68f9698c06326fab5ab837701e11a1315e158e46888eb81c2db09ea62`

Algorithm	Hash digest
SHA256	`a7bb41bbd1cb11dec0252caf0e6ab76fabdcaae9788fb1aebd733d4617165da1`
MD5	`04cfd98a2b4ec254f5ba57cfda6a27a7`
BLAKE2b-256	`98a0babc9828cc23a41ab691d7cfe5c260c7c9b80d72615898b0c80d5da57bde`

Algorithm	Hash digest
SHA256	`155a50292286787f86f361fd0066b4ac4acc156db17cc7ae8c6751ae84eb8ed5`
MD5	`b64be9cc94823d80f907cb5b014d51f8`
BLAKE2b-256	`985197d25a00c01920d2df7f4a17dfbfde46c80086ef123f6ddaa6a53008ffd8`

Algorithm	Hash digest
SHA256	`a772a36ad8194931aa69b79421cc260ea6b3189ac37e14fa44be993fd9e1db45`
MD5	`990d5d27cb1bfa1e23d13df5e1dc95c4`
BLAKE2b-256	`f3ef827e984c61136f5e343289c984af829bd79cd56916d58cb18bdeaefb5e9a`

Algorithm	Hash digest
SHA256	`f801bb21137e24247d99ea1d45eaccf94de93e4a3332897abbffba015172b5d6`
MD5	`ef9aeea57e150a825b7090fbc5852dc9`
BLAKE2b-256	`1a2bc6b0c84deec7bb491406c013ac059df955b3221ad0f6d08ae0987ffca24f`

Algorithm	Hash digest
SHA256	`0fbcdcd047b1f33ef5cd415f21dd20a191069d1e7864457a4b4efa36de1e27c2`
MD5	`ad42c8662f67d8f5470c801639047781`
BLAKE2b-256	`9f94e3a2fabfa9c8561e7cd1cbee5ccca1c9d273df64d4d1b1d563c9793e2605`

Algorithm	Hash digest
SHA256	`08d217aa1de0376fe63f0ad6b56b6ddf047efbfcb8dfbddb9f3beaed605b9e81`
MD5	`29bb751aaf367b368cc9aad1c0c472a2`
BLAKE2b-256	`3cf267dba161e35b80835bbebe754575b696945602ec298f0947a97996bd780d`

Algorithm	Hash digest
SHA256	`1a26e02045e38be2cda0ca88012a13dc5cead6c7a7f7167d0540f136e4b8e993`
MD5	`0935b047745aa159dbfb1b932bc7a935`
BLAKE2b-256	`63a4d8336d68cae54d9cc53e25d96cfcac0af3d1e2251e3060e281655b1ba180`

Algorithm	Hash digest
SHA256	`a632155d7979315a6bc9ff0c34909665924d775ba4a4f5557c92cf023ef633a7`
MD5	`2034730e2ad399bc8b03e9f321b40445`
BLAKE2b-256	`2aece28c57c22a7399cb977371db193d8a3cfe80a4db17a6f3e17101a3da65b9`

Algorithm	Hash digest
SHA256	`0eddcba8970fbcb9ebbaf88a3e5a4ebe9b9229d12c6e772c85ff62458764b50c`
MD5	`4151608eee66ba1f67de992a9df91b09`
BLAKE2b-256	`b87b9c1c9e39bc2842eb0fcbd1d9aaaf950e4cad5d1529e7f97478187f5a2903`

Algorithm	Hash digest
SHA256	`4c6e10976a235fd2a04f99e8ce4e000eef453afc2fbdbe495981d878d3da7833`
MD5	`5566e56b09fa9f2af0ef15a2fe972ef2`
BLAKE2b-256	`b9ac53a7e0baf423355d2cff30c71771f15c59f08b7b9631f5d32c1d3ea108ab`

Algorithm	Hash digest
SHA256	`8668680a4b7763dfc8c23c5970cb54af5ecdef5f1acb9d3e133e54c0cb30dac2`
MD5	`cb73cd717ded98b89d613d392f8f9321`
BLAKE2b-256	`01fd5f2366910e751bde90498cabf5ed9c10086cf86a094f41986620a4e6fe9d`

ringgrid 0.11.0

ringgrid (Python)

Install

Fast Start: Generate target_spec.json + Printable SVG/PNG/DXF

Target layouts

Features

Input Rules

Proposal-Only Diagnostics

DetectConfig Field Guide

Surface Map

Top-Level Controls

Convenience Aliases

marker_scale

proposal

edge_sample

outer_estimation

marker_spec

outer_fit

inner_fit

decode

seed_proposals

projective_center

completion

ransac_homography

self_undistort

id_correction

inner_as_outer_recovery

Adaptive Detection

Which Method Should I Use?

Practical Recipes

Examples

Project links

Key dates

1 maintainer

Credits

GitHub Statistics

License

Requires

Provides Extra

Classifiers

Download files

Source Distributions

Built Distributions

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

File metadata

File hashes

Provenance

File details

Fast Start: Generate `target_spec.json` + Printable SVG/PNG/DXF

`marker_scale`

`proposal`

`edge_sample`

`outer_estimation`

`marker_spec`

`outer_fit`

`inner_fit`

`decode`

`seed_proposals`

`projective_center`

`completion`

`ransac_homography`

`self_undistort`

`id_correction`

`inner_as_outer_recovery`