counting op• 数据种类:image2d → table
• 调用:import photoncount; photoncount.photon_statistics(counts)(或 opsphoton.get("photon_statistics"))
光子计数帧的泊松统计:它真的是散粒噪声受限的吗?
> 以下的详细说明为原文 —— 摘要与标题已翻译。
Returns a dict: `mean · variance (population, ddof=0`) ·
`fano_factor = variance / mean` (1 for a Poisson process) ·
`snr_poisson = sqrt(mean)` (the theoretical photon-limited SNR) ·
`snr_measured = mean / std` (what this frame actually achieved) ·
`total_counts · n_samples · zero_fraction` (the fraction of pixels
that saw no photon at all — the honest measure of "photon starved";
`exp(-lambda) for a flat field) · max_counts`.
The Fano factor is evidence of Poisson statistics only on a flat field.
On a structured scene the scene's own spatial variance dominates and the
ratio is large and meaningless — this op computes the number, it cannot tell
you which situation you are in. Measured on the test scenes: a flat
`lambda = 100` field (512x512, seed 0) gives 1.001089; the same detector
looking at a linear ramp from 20 to 180 photons gives 22.4102. Both are
"correct" and only one of them means anything.
Raises `ValueError`: negative, non-finite or non-2-D *counts*, fewer
than 2 pixels (no variance), an all-zero frame (`fano_factor` would be
`0/0` — say "no photons were detected" instead of returning NaN), and a
frame with exactly zero variance (`snr_measured would be inf`; for
`n >= 2` a constant frame is not a Poisson realisation but a synthetic
constant, i.e. an input mistake).
• 示例数据目录(下载 URL / 许可证) —— 2-D 用 skimage.data(BSD/公有领域)加合成图,3-D 给出真实数据源(Stanford/PDS 等)的下载 URL。
• 算子来历与参考文献 —— 该算子族所依据的研究/方法出处。
• 算法的正典(作者・年份)与用途见上面的族使用指南。
• photon_timeresolved — py -3.11 examples/photon_timeresolved.py
table 作为输入)—
counting)photon_sample · photon_uncertainty
*Provenance: photoncount.py — PHOTON 算子登记表。本条目由 tools/opdocs.py md 自动生成(请勿手工编辑)。*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.