tracking op• Data kinds: none → table (an op determined by its arguments alone — it takes no image or data input)
• Call: import fullseye as fs; fs.ledger.filter_kalman(x, P, z, F, H, Q, R) (to call the implementation directly, import mvtrack; mvtrack.filter_kalman(x, P, z, F, H, Q, R); from the registry, opsmv.get("filter_kalman"))
• HALCON equivalent: filter_kalman (the HALCON reference is a useful guide to its meaning and parameters)
• GPU: none (CPU reference implementation only)
Run one predict-and-update cycle of a linear Kalman filter.
> The detailed description below is the original text — the summary and the headings are translated.
x' = F x、P' = F P F^T + Q で予測し、観測 z で更新する。共分散は対称性の崩れに強い Joseph 形で更新する。z が None なら更新を省く(観測の欠けたコマ)。空の入力(x が 0 要素)は ValueError。
Args:
x: 状態 (n,)。
P: 状態の共分散 (n, n)。
z: 観測 (m,)。None で予測だけ。
F: 遷移行列 (n, n)。
H: 観測行列 (m, n)。
Q: 過程の雑音の共分散 (n, n)(スカラーは対角へ広げる)。
R: 観測の雑音の共分散 (m, m)(スカラーは対角へ広げる)。
Returns:
x、P、gain、innovation、nis(正規化した新息の 2 乗)を持つ dict。
観測が無いコマの gain と innovation は 0、nis は NaN。
• machine_vision_classics family guide
• Sample-data catalog (download URLs / licences) — 2-D uses skimage.data (BSD/public domain) plus synthetic images; 3-D lists download URLs for real data sources (Stanford, PDS, …).
• Operator provenance and references — the sources of the research/methods this op family came from.
• The canonical algorithm (author, year) and its uses are named in the family usage guide above.
• machine_vision_tour — py -3.11 examples/machine_vision_tour.py
table as input)map_image · rectify_to_world_plane · change_radial_distortion_image · find_ncc_models · run_bg_esti · give_bg_esti · update_bg_esti · classify_class_gmm
tracking)create_bg_esti · run_bg_esti · give_bg_esti · update_bg_esti
*Provenance: mvtrack.py — MV operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.