typed opsignal → signalfullseye.apply(img, "tb_fly_tau_from_expansion", a=0.5, b=0.5) (2-D は 1 画像 + 2 スカラつまみ a,b∈[0,1] のモデル)
図は合成の入力 128×128 で実際に走らせた出力。左が入力、右が出力。点群は上から見た散布(明るさ = z)、1-D 列は折れ線、体積は z 方向の最大値投影、動画は中央フレーム、複素画像は振幅、絵にならない返り値は値そのもの。
つまみ a は出力を変えない(実測: 0.1 / 0.5 / 0.9 で同一)。
つまみ b は出力を変えない(実測: 0.1 / 0.5 / 0.9 で同一)。
段階(前置きの op → この op。左から順):

別の画像でも(合成シーン / 写真 / 硬貨。上段が入力、下段がその出力。つまみは既定):

Time-to-contact from optical expansion — the tau margin.
From the subtended angle and its rate::
shape="disk": tau = sin(theta) / theta'
shape="sphere": tau = 2*tan(theta/2) / theta'
theta_signal is the full subtended angle over time, radians.
★ The two object models differ by 33% at a 60-degree subtense (sin 60 = 0.866
vs 2 tan 30 = 1.155); using the wrong one silently mis-times a landing, so
the model is a named choice, with "sphere" as the default.
theta_signal: the full subtended angle per sample, radians.
dt_s: the sample interval, seconds.
shape: "disk" (a frontal circular disk) or "sphere".
Returns a 1-D float64 array of the time-to-contact per sample, seconds.
Non-expanding samples (theta' <= 0) return NaN — a documented
non-finite, because a contracting or static angle has no time-to-contact and
inventing one would be a plausible-wrong number. (The registry op
tb_fly_tau_from_expansion must return a finite signal, so it replaces those
NaN by the signal fallback without recording a fallback event —
backend_safe.NONFINITE_BY_DESIGN; call this function directly to keep the NaN.)
Ground truth: for the model’s own object geometry (theta = 2 asin(l/d) for a
sphere, theta = 2 atan(l/d) for a disk) approaching at speed |v|, the
returned tau equals the true distance-over-speed d/|v| (pinned in the
tests).
Raises ValueError: a non-1-D / empty / too-short (< 3) / non-finite
theta_signal, a signal over :data:MAX_SIGNAL_POINTS, a non-positive dt_s,
and an unknown shape. A non-expanding angle is not an error — it is the
documented NaN return above.
Typed bridge of the flyvision op fly_tau_from_expansion into the 2-D evolution registry: the same implementation, called under the op(v, a, b) convention. This op has no tunable parameter; a and b are unused.
下のプログラムは実際に走ることを確かめてある(図と同じ入力)。Studio のヘルプではこのブロックがボタンになり、その場で読み込んで実行できる。
img_to_signal 0.50 0.50
tb_fly_tau_from_expansion 0.50 0.50
次の例は元の台帳 op fly_tau_from_expansion を呼ぶもの。この橋渡し op は同じ実装を fn(v, a, b) 規約に合わせただけなので、挙動はそのまま当てはまる(呼び出し形だけ違う)。
py -3.11 examples/poc_fly_vision.pypy -3.11 examples/poc_zebra_stripes_fly_eye.pysignal を入力に取れる)identity · tb_create_funct_1d_array · tb_smooth_funct_1d_gauss · tb_smooth_funct_1d_mean · tb_derivate_funct_1d · tb_integrate_funct_1d · tb_zero_crossings_funct_1d · tb_abs_funct_1d
typed)tb_points_to_voxel · tb_estimate_point_normals · tb_iss_keypoints · tb_project_points · tb_render_point_depth · tb_points_render · tb_statistical_outlier_removal · tb_radius_outlier_removal
Provenance: ops.py — 2D 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.