calibration_error_map — CALIB calib_truth op

• Data kinds: table → table

• Call: import fullseye as fs; fs.ledger.calibration_error_map(calibration, K_true, dist_true, image_shape, step=16, depth=None, remove_rotation=True) (to call the implementation directly, import calib_lm; calib_lm.calibration_error_map(calibration, K_true, dist_true, image_shape, step=16, depth=None, remove_rotation=True); from the registry, opscalib.get("calibration_error_map"))

• GPU: none (CPU reference implementation only)

Usage

Compares the true K and distortion with the estimate and returns a map of where in the image the reprojection error (px) and 3-D error (mm) are large (table).

> The detailed description below is the original text — the summary and the headings are translated.

Map, over the image, how far the estimated camera misplaces a point compared with the true camera — possible only when the truth is known (a simulated calibration). For each grid pixel p the ray of the TRUE camera through p is reprojected with the ESTIMATED camera; the map is |p̂ − p| in pixels (the error any reprojection-based measurement inherits there). With depth Z (mm), the 3-D map is the lateral distance between the points the two cameras back-project from p at depth Z.

remove_rotation(既定 True): 推定のカメラを最もよく合う小さな回転で回してから比べる。主点のずれの大部分はカメラの向きのずれと見分けがつかず、どんな使い方でも外部の姿勢に吸収されるので、残るのが「内部パラメータと歪みでしか直せない」誤差になる。回した角度は rotation_removed_deg、回す前の地図は raw_* に残す。閉形式の門(remove_rotation=False): 推定 = 真値 → 0、cx だけ δ ずれ(歪みなし)→ どこでも δ px、3D は δ·Z/fx。image_shape (H, W)、step は格子の間隔(画素)。

返す dict: rows・cols(格子の中心)、reproj_error_px(地図)、dx_px・dy_px(成分、列・行)、max_px・rms_px・max_at(row, col)、raw_rms_px・raw_max_px(回す前)、rotation_removed_deg、error_3d_mm・max_3d_mm(depth があれば)、depth。

Raises ValueError: 校正の表が不正、K が正でない、image_shape・step・depth の範囲外。

Detailed usage guide

• camera_calibration family guide

References (sample data, literature)

• Sample-data catalog (download URLs / licences) — 2-D uses skimage.data (BSD/public domain) plus synthetic images; 3-D lists download URLs for real data sources (Stanford, PDS, …).

• Operator provenance and references — the sources of the research/methods this op family came from.

• The canonical algorithm (author, year) and its uses are named in the family usage guide above.

Runnable examples (verified samples that actually call this op)

• poc_calibration_truth_map — py -3.11 examples/poc_calibration_truth_map.py

Ops the type connects to (they accept table as input)

image_points_to_world_plane · contour_to_world_plane_xld · sim_caltab · disp_caltab · calibration_uncertainty_map · calibration_next_pose

Same category (calib_truth)

calibration_uncertainty_map · calibration_next_pose


*Provenance: calib_lm.py — CALIB operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*

© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.