scene op• Data kinds: table × image2d → table
• Call: import optscene; optscene.surface_defect(primitive: 'dict', field, mask=None, uv_size_mm=(20.0, 20.0), centre_mm=(0.0, 0.0), height_um: 'float' = 0.0, height_field=None, roughness_um: 'float' = 0.6) -> 'dict' (or opsoptics.get("surface_defect"))
Paste a 2-D defect map onto a part's surface (it eats the output of `defectgen` directly).
> The detailed description below is the original text — the summary and the headings are translated.
`field` は明るさの変調 (H, W)(0 = 変化なし、−0.3 = 30% 暗い傷)。
`height_um` を与えると同じ図を高さとも解釈し、勾配から法線を傾ける
―― これがあると、同じ傷がドーム照明では消え暗視野照明で光る、という
照明を変える意味が再現される(外観検査 AI の学習データはここが本体)。
`height_field` を別に渡すと、色は変わらないが凹凸だけがある欠陥(打痕・
ひけ・浅い擦り傷)を作れる。`field` を全ゼロにすれば純粋な地形欠陥になり、
ドーム照明では消えて低角の暗視野照明で光る ―― この差こそ照明を選ぶ理由。
`roughness_um` は欠陥のところだけ面が粗くなる量(既定 0.6 µm)。
見方を変えると、:func:surface_finish の加工目も同じものを別の粒度で持っている
―― 加工目は表面凹凸のテクスチャを数オクターブの雑音に圧縮した表現で、
ここでいう粗さは「画素より細かくて解像できない凹凸」を 1 つの数(σ)に潰した表現。
解像できる凹凸は法線として、解像できない凹凸は粗さとして扱う、という分担である。傷は材料を
削り取った跡なので、健全面より必ず粗い。これがあると
鏡面割合 exp(-(4πσcosθ/λ)²) が下がって明視野で暗くなり、同時に散乱が増えて
暗視野で明るくなる ―― 教科書どおりのコントラスト反転は、法線の傾きではなく
この粗さから出る(低角の光を真上へ返すには面が 39 度傾く必要があり、傷の傾斜では
届かないため。2026-09-05 に実測で確認)。0 にすれば粗さを変えない欠陥になる。
`uv_size_mm は貼り付ける実寸 [mm]、centre_mm` は面座標上の中心。
`mask を渡すとその画素が欠陥ラベル(optscene_defect_mask` が返す真値)。
返り値は defect を付けた新しいプリミティブ(元は書き換えない)。
Every optics op validates its input before computing (nothing slips through silently):
• Units are baked into the argument name — _mm / _um / _deg / _mrad. Confusing mm with µm does not crash; it yields a plausible-looking wrong answer, so the name prevents it. Nothing here guesses the unit from the magnitude.
• **Strings raise ValueError** — float('50') succeeds, so an unparsed configuration value would slip through as a length (measured: thin_lens('50', '200') returned a plausible 66.667 mm). bool is refused too, as the implicit promotion True == 1.
• **complex / masked arrays raise ValueError (real-valued slots only; silently dropping the imaginary part or peeling off the mask is refused). NaN/Inf raises ValueError on every input.**
• Division by zero and its relatives are refused by name: focal length 0, radius of curvature 0, refractive index <= 0, a fully opaque aperture (all zeros, so the normalisation is 0/0), a PSF whose sum is <= 0, a Stokes vector with S0 = 0, and an object sitting at the front focal point (the image is at infinity).
• Only two ops return a non-finite value, and both state it as a contract: depth_of_field returns far_mm = inf beyond the hyperfocal distance (that is what the hyperfocal distance means), and gaussian_beam returns wavefront_radius_mm = inf at the waist (the radius of curvature of a plane wavefront). Both also return a finite companion (far_is_infinite / curvature_per_mm). **Any other silent NaN/Inf is detected internally and raises ValueError** — "float64 overflowed" and "the answer is infinite" are different claims, so the first is never returned wearing the face of the second.
• Size caps: generated grids are capped by optics.MAX_GRID (4096); supplied fields/PSFs/apertures by optics.MAX_FIELD_ELEMENTS (2^24); ABCD element chains by optics.MAX_SYSTEM_ELEMENTS (1024); Zernike by MAX_ZERNIKE_TERMS (512) / MAX_ZERNIKE_ORDER (40) / MAX_ZERNIKE_BASIS (2^25). This closes, fail-closed, the paths where a small argument triggers a huge internal allocation (measured: n_max=40 × 4096² needs 108 GB).
• Physically impossible states are refused too: a Stokes vector with degree of polarisation > 1, negative transmittance, negative intensity, and invalid Zernike indices such as n-|m| odd.
• dataset_conventions — 学習データセット規約の知識 — COCO / YOLO / VOC と外観検査での落とし穴
• mv_cameras — 産業用カメラメーカー（センサとの紐付け・ラインスキャン / TDI）
• mv_illumination_practice — 照明の実務知識 — 波長・偏光・点灯方式・外光・安全
• mv_image_sensors — 産業用イメージセンサ（現行品中心）
• virtual_machine_vision — 仮想マシンビジョン — パラメータの洗い出しとオブジェクト模型
• 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.
• studio_raytrace_scene — py -3.11 examples/studio_raytrace_scene.py
table as input)abcd_matrix · wavefront_stats · paraxial_trace · seidel_coefficients · spot_stats · tolerance_analysis · wavefront_from_opd · spot_diagram
scene)scene_material · scene_plane · scene_sphere · scene_box · scene_cylinder · surface_finish · random_defects · scene_difference
*Provenance: optscene.py — OPTICS 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.