fullseye

tb_fly_tau_from_expansion — 2D typed op

tb_fly_tau_from_expansion: input → output

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):

tb_fly_tau_from_expansion: stages

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):

tb_fly_tau_from_expansion: other inputs

使い方

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 required, named choice rather than a default guess.

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.

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.

参考(サンプルデータ・文献)

Studio で試す

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

実行できる例(この op を実際に呼ぶ検証済みサンプル)

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)。

型が繋がる次の op(signal 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

同カテゴリ(typed)

tb_points_to_voxel · tb_estimate_point_normals · tb_iss_keypoints · tb_project_points · tb_render_point_depth · tb_statistical_outlier_removal · tb_radius_outlier_removal · tb_voxel_grid_downsample


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.