tracktable.algorithms.distance_geometry module

Module contents

tracktable.algorithms.distance_geometry - Compute the distance geometry signature of a trajectory.

tracktable.algorithms.distance_geometry.distance_geometry_by_distance(trajectory, depth)[source]

Compute distance geometry signature sampled by length

This function computes the multilevel distance geometry for a given trajectory. Each level d approximates the input trajectory with d equal-length line segments. The distance geometry values for that level are the lengths of all d line segments, normalized to lie between 0 and 1. A value of 1 indicates the length of the entire trajectory.

The D-level distance geometry for a curve will result in (D * (D+1)) / 2 separate values.

This implementation creates the endpoints of the line segments by sampling the trajectory at fractions of total distance traveled. To sample by total duration, use distance_geometry_by_time().

Parameters

  • trajectory (Tracktable trajectory) – Input curve to analyze

  • depth (int) – How many levels to compute. Must be greater than zero.

Returns

List of distance geometry values laid out consecutively by increasing depth.

Raises

  • ValueError – depth is not a positive integer

  • BoostPythonArgumentException – trajectory is not a Tracktable trajectory

tracktable.algorithms.distance_geometry.distance_geometry_by_time(trajectory, depth)[source]

Compute distance geometry signature sampled by time

This function computes the multilevel distance geometry for a given trajectory. Each level d approximates the input trajectory with d equal-length line segments. The distance geometry values for that level are the lengths of all d line segments, normalized to lie between 0 and 1. A value of 1 indicates the length of the entire trajectory.

The D-level distance geometry for a curve will result in (D * (D+1)) / 2 separate values.

This implementation creates the endpoints of the line segments by sampling the trajectory at fractions of total duration To sample by distance traveled, use distance_geometry_by_distance().

Parameters

  • trajectory (Tracktable trajectory) – Input curve to analyze

  • depth (int) – How many levels to compute. Must be greater than zero.

Returns

List of distance geometry values laid out consecutively by increasing depth.

Raises

  • ValueError – depth is not a positive integer

  • BoostPythonArgumentException – trajectory is not a Tracktable trajectory