spad op• Datenarten: counts → counts
• Aufruf: import photoncount; photoncount.tcspc_coates_correct(hist, cycles) (oder opsphoton.get("tcspc_coates_correct"))
Macht TCSPC-Pile-up exakt rückgängig (Coates-Schätzer) — die Frühphotonen-Verzerrung.
> Die ausführliche Beschreibung unten ist der Originaltext — Zusammenfassung und Überschriften sind übersetzt.
Classical TCSPC records at most one photon per excitation cycle: the
first one. Late bins are therefore starved, because the cycles in which an
early photon arrived never reach them, and the measured histogram is biased
toward short arrival times — a dToF depth read straight off a piled-up
histogram is *too close*, and a fluorescence lifetime is *too short*.
Coates's estimator inverts that exactly. With `N_k` the measured counts in
bin `k and C` the number of excitation cycles, the number of cycles that
survived to reach bin `k is D_k = C - sum_{j<k} N_j`, the per-cycle
detection probability in that bin is `p_k = N_k / D_k` and the pile-up-free
per-cycle intensity is `lambda_k = -ln(1 - p_k)`. This op returns
`C * lambda_k` — the histogram the same scene would have produced if the
detector could record every photon — so it is directly comparable to the
measured one.
This is an exact inverse, not a linearisation: build a histogram from a
known `lambda through the forward model `N_k = C * exp(-sum_{j<k}
lambda_j) * (1 - exp(-lambda_k))` and Coates returns lambda` to machine
precision (measured max relative error 1.6e-15 in the tests, on a pile-up so
severe that the last bin was suppressed to 14.8% of its true counts).
*hist* is the 1-D measured histogram (counts per bin); *cycles* the number of
excitation cycles (laser pulses) that produced it.
Raises `ValueError`: negative, non-finite or non-1-D *hist*, a
non-positive or non-integer *cycles*, a histogram whose total exceeds
*cycles* (impossible: at most one photon per cycle — a sure sign that
*cycles* is wrong or the data are not first-photon TCSPC), and any bin that
consumed every remaining cycle (`p_k = 1, where -ln(0) is inf`).
• Leitfaden zur Familie photon_timeresolved
• Katalog der Beispieldaten (Download-URLs / Lizenzen) — 2-D nutzt skimage.data (BSD/Public Domain) plus synthetische Bilder, 3-D nennt Download-URLs echter Datenquellen (Stanford, PDS, …).
• Herkunft und Literatur der Operatoren — die Quellen der Forschung/Verfahren, auf denen diese Operatorfamilie beruht.
• Der kanonische Algorithmus (Autor, Jahr) und seine Anwendungen stehen im Familienleitfaden oben.
• photon_timeresolved — py -3.11 examples/photon_timeresolved.py
counts als Eingabe)tcspc_irf_convolve · tcspc_background_subtract · tcspc_stats · dtof_depth · lifetime_fit · lifetime_phasor
spad)spad_deadtime_apply · spad_deadtime_correct
*Provenance: photoncount.py — PHOTON Operator-Registry. Diese Notiz wird von tools/opdocs.py md erzeugt (nicht von Hand bearbeiten).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.