depth op• 데이터 종류: image2d → depth
• 호출: import lightfield; lightfield.lf_disparity_to_depth(slope, focal_px=1000.0, baseline=1.0, *, far_depth=None, min_slope=1e-06)(또는 opslightfield.get("lf_disparity_to_depth"))
기울기(각도 한 단계당 화소) -> 실제 깊이. 카메라 어레이 모델.
> 아래 상세 설명은 원문입니다 —— 요약과 제목은 번역되어 있습니다.
For a rectified array of viewpoints spaced *baseline* apart with focal
length *focal_px* expressed in pixels of the sub-aperture image, a point
at distance `Z shifts by |s| = focal_px * baseline / Z` pixels per
step. Inverting: `Z = focal_px * baseline / |s|`, returned in whatever
length unit *baseline* was given in (mm in, mm out).
Only `|s|` enters, deliberately. The *sign* of the slope says which side of
the focal plane a point is on, and which sign means "nearer" depends on how
the angular axis was oriented when the field was decoded — a convention this
module cannot know and refuses to guess. If your decode puts near objects at
negative slope, negate before calling.
*far_depth* is the fail-closed switch for the `s -> 0` pole (a point at
infinity has zero parallax). `None (default) makes any |s| < min_slope`
a `ValueError` naming how many pixels were affected — mask them, or pass an
explicit saturation distance as *far_depth* and get that value there. There
is no silent `inf` and no silent clamp.
Accepts a scalar or any array; returns the same shape (a scalar in gives a
0-d array out, so downstream code has one type to handle).
Raises `ValueError`: non-finite *slope* / *focal_px* / *baseline*, a
non-positive *focal_px* / *baseline* / *min_slope*, a negative *far_depth*,
and `|slope| < min_slope anywhere when *far_depth* is None`.
• 샘플 데이터 카탈로그(DL URL / 라이선스) —— 2-D 는 skimage.data(BSD/public)+ 합성, 3-D 는 실데이터 소스(Stanford/PDS 등)의 DL URL.
• 연산자의 내력·참고문헌 —— 이 연산자 족의 바탕이 된 연구/기법의 출처.
• 알고리즘의 정전(저자·연도)과 용도는 위의 패밀리 사용 가이드에 적혀 있습니다.
• lightfield_depth — py -3.11 examples/lightfield_depth.py
depth 를 입력으로 받는 것)—
depth)lf_depth_from_focus · lf_epi_slope · lf_all_in_focus · lf_plenoptic_design
*Provenance: lightfield.py — LIGHTFIELD 연산자 레지스트리. 이 op 노트는 tools/opdocs.py md 가 자동 생성합니다(직접 편집하지 마세요).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.