ObjectEstimate

class rapidtools.processing.street_tracking.ObjectEstimate(x: float, y: float, cov: ~numpy.ndarray, members: list[~rapidtools.core.observation.Observation], localization: str, rms_m: float | None = None, parallax_deg: float = 0.0, rays: list[tuple[float, float, float]] = <factory>, size_m: float | None = None)

Bases: object

One candidate object: a position with uncertainty and its sightings.

Variables:

• x (float) – East coordinate in the local metre frame.

• y (float) – North coordinate in the local metre frame.

• cov (numpy.ndarray) – 2 x 2 position covariance in square metres.

• members (list[Observation]) – The sightings that produced the estimate.

• localization (str) – 'triangulated' (rays intersected) or 'single_view' (ground-contact ranges only).

• rms_m (float | None) – Root-mean-square ray residual of a triangulation, else None.

• parallax_deg (float) – Largest bearing separation among the rays used.

• rays (list[tuple[float, float, float]]) – (camera_x, camera_y, bearing_deg) per frame, the geometry that is reliable regardless of range; used to associate fragments and duplicates by their lines of sight.

• size_m (float | None) – Physical width of the object as seen (median over frames), or None when unknown.

property image_ids: set[str]

Frames that saw the object.

property sequence_ids: set[str]

Sequences that saw the object.

property n_images: int

Number of distinct frames.

property sigma_m: float

Largest one-sigma extent of the position uncertainty in metres.