balltrack op• Data kinds: any → table
• Call: import fullseye as fs; fs.ledger.spin_from_marker_sequence(dirs, dt: 'float', *, lag: 'int' = 4, max_angle: 'float' = 0.5) -> 'dict' (to call the implementation directly, import balltrack; balltrack.spin_from_marker_sequence(dirs, dt: 'float', *, lag: 'int' = 4, max_angle: 'float' = 0.5) -> 'dict'; from the registry, opsdrive.get("spin_from_marker_sequence"))
Reads the angular velocity from the marker directions dirs[k] over a sequence of frames ((M_k, 3) unit vectors; None for frames where none is found) (2 stages).
Stage 1: between adjacent frames, pair the closest markers whose directions are within max_angle [rad] and apply Kabsch (spin_from_markers); the median is ω₁. Stage 2: pairs lag frames apart are refitted by pairing each with the marker closest (within max_angle/2) to the prediction obtained by rotating the earlier frame by ω₁ —— since the rotation angle is lag times larger, the relative effect of marker position error is 1/lag. ★The rotation angle per frame must be well below max_angle (otherwise markers get mixed up), and the rotation over lag frames must be less than π (Kabsch returns the shortest rotation). Return value {"omega", "omega_stage1", "n_pairs", "per_pair" (N, 3)} (ω is in the same coordinate system as dirs). If there is not a single pair, ω is NaN. Pairs of directions where markers were seen are needed, so each frame needs 2 or more markers.
Empty input (dirs with 0 entries) raises ValueError.
• 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.
• poc_table_tennis_bounce — py -3.11 examples/poc_table_tennis_bounce.py
• poc_table_tennis_spin — py -3.11 examples/poc_table_tennis_spin.py
table as input)course_layout · course_occupancy · course_contains · world_build · world_camera · world_move · lidar_scan · rss_longitudinal_same
balltrack)ball_detect · ball_track · kalman_ca · triangulate_dlt · track_triangulate · bounce_detect · marker_direction · spin_from_markers
*Provenance: balltrack.py — DRIVE 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.