vote_rays

Contents

vote_rays#

rapidtools.processing.street_tracking.vote_rays(observations: Sequence[Observation], project: Callable[[float, float], tuple[float, float]], min_range_m: float = 2.0, max_range_m: float = 30.0, cell_m: float = 0.5, object_size_m: float = 4.5, min_score: float = 1.5) → list[ObjectEstimate][source]#

Locate objects as peaks of accumulated ray votes on a ground grid.

Every sighting spreads one unit of weight along its ray between min_range_m and max_range_m, concentrated around its ground-contact range when it has one. Objects are local maxima of the smoothed grid, separated by at least object_size_m; each sighting is then attributed to the nearest peak along its ray.

Parameters:

  • observations – Sightings with bearing set.

  • project – Local metre projection.

  • min_range_m – Nearest range a vote can land at.

  • max_range_m – Farthest range a vote can land at.

  • cell_m – Grid resolution.

  • object_size_m – Non-maximum-suppression window.

  • min_score – Smallest peak that counts as an object; about the number of sightings that must agree.

Returns:

list[ObjectEstimate] – One estimate per peak, localization 'voted'.

Contents