fullseye

dtof_cube_simulate — PHOTON dtof op

• データ種: depth → histcube
• 呼び出し: import fullseye as fs; fs.ledger.dtof_cube_simulate(depth, bins=256, bin_ps=100.0, reflectivity=None, signal_photons=20.0, ambient_photons=5.0, irf_fwhm_ps=200.0, seed=0, noise=True) (実装を直接呼ぶなら import photoncount; photoncount.dtof_cube_simulate(depth, bins=256, bin_ps=100.0, reflectivity=None, signal_photons=20.0, ambient_photons=5.0, irf_fwhm_ps=200.0, seed=0, noise=True)、台帳から引くなら opsphoton.get("dtof_cube_simulate"))

使い方

Synthesise the (H, W, T) photon histogram cube a SPAD array produces.

The per-pixel version of :func:tcspc_simulate: every pixel of the depth map (metres, one-way distance) gets a Gaussian return at its own round-trip time 2d/c, scaled by signal_photons times that pixel’s reflectivity, on a uniform ambient pedestal of ambient_photons/bins per bin, Poisson sampled with numpy.random.default_rng(seed).

The output is the cube that :func:dtof_cube_depth inverts, and the axis order is (H, W, T) with time LAST — the same layout a SPAD array streams. That is not the (D, H, W) of a :mod:volops voxel volume; the two are both 3-D float arrays and swapping them silently produces a plausible-wrong depth map, which is why :func:dtof_cube_depth checks and says so.

noise=False returns the exact expectation cube (no sampling).

Ground truth: with noise=False the per-pixel centroid of the cube returns the input depth map to an RMS error of 3.2e-16 m (pinned in the tests) — the pulse integral is analytic, so the only error is float round-off.

Raises ValueError: a non-2-D, non-finite or non-positive depth, a reflectivity that is negative or not the same shape as depth, a bins outside [2, MAX_BINS], non-positive bin_ps / irf_fwhm_ps, negative photon budgets, a non-integer seed, a cube over :data:MAX_CUBE_ELEMENTS (H*W*bins grows fast — 512x512x256 is 8x the cap), and any depth whose round-trip time falls outside the time window (which a real sensor would alias into a short distance).

詳しい使い方ガイド

• photon_timeresolved ファミリ ガイド

参考(サンプルデータ・文献)

• サンプルデータ カタログ(DL URL / ライセンス) — 2-D は skimage.data(BSD/public)+ 合成、3-D は実データ源(Stanford/PDS 等)の DL URL。
• 演算子の来歴・参考文献 — この op 族の元になった研究/手法の出典。
• アルゴリズムの正典(著者・年)と用途は上記ファミリ使い方ガイドに記載。

実行できる例(この op を実際に呼ぶ検証済みサンプル)

• photon_timeresolved — py -3.11 examples/photon_timeresolved.py
• poc_dtof_ranging — py -3.11 examples/poc_dtof_ranging.py

型が繋がる次の op(histcube を入力に取れる)

dtof_cube_depth

同カテゴリ(dtof)

dtof_depth · dtof_cube_depth

---

Provenance: photoncount.py — PHOTON operator registry. この per-op ノートは tools/opdocs.py md が自動生成(手編集しない)。

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

This site is open source. Improve this page.