typed opbeatcube → imagefullseye.apply(img, "tb_range_doppler_map", 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):

The 2-D FFT of a beat cube -> a (n_doppler, n_range) magnitude map.
Fast time transforms to **range** (last axis, not shifted: bin ``j`` is
``j * c*f_s/(2*S*N_s)`` metres, and a physical range is always positive so
the whole ``[0, f_s)`` band is used). Slow time transforms to **velocity**
(middle axis, ``fftshift``ed so the map is centred on zero velocity: bin
``i`` is ``(i - N_c//2) * lambda/(2*N_c*T_c)`` metres per second, positive =
receding).
The antenna axis is collapsed by *combine*: ``"incoherent"`` (default) is the
**mean of the magnitudes**, which is angle independent and therefore the
right default for detection; ``"coherent"`` is the **magnitude of the mean**,
i.e. a beam pointed at boresight, which attenuates an off-boresight target on
purpose. ``antenna=k`` uses element ``k`` alone. For a single-element cube
all three agree exactly.
``normalize=True`` divides by ``N_c * N_s``, so a bin-centred target of
amplitude ``a`` peaks at exactly ``a`` (measured: 1.0 for a unit target,
absolute error 0.0). The default ``False`` keeps the raw FFT magnitude.
No window is applied — compose :func:`fmcw_window_apply` first if you want
one. The output is a plain 2-D float64 array, so every 2-D operator in
Fullseye (threshold, morphology, labelling, blob measurement — the pieces a
CFAR detector is made of) applies to it directly.
**Raises** ``ValueError``: a real-valued cube (it would put a mirror ghost of
every target at a fabricated range), fewer than 2 chirps or 2 samples, an
out-of-range *antenna* index, an unknown *combine*, a cube over the element
cap, an FFT that overflows to NaN, or NaN/Inf on the way in.
Typed bridge of the rangedoppler op range_doppler_map 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_beatcube 0.50 0.50
tb_range_doppler_map 0.50 0.50
次の例は元の台帳 op range_doppler_map 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/fmcw_range_doppler.pyimage を入力に取れる)identity · gaussian · mean_box · bilateral · unsharp · median · min_filter · max_filter
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.