cars.applications.ground_truth_reprojection.ground_truth_reprojection_tools

This module contains the abstract direct_localization application class.

Module Contents

Functions

get_ground_truth(geom_plugin, grid, sensor, ...[, ...])

Computes ground truth in epipolar and sensor geometry.

resample_auxiliary_values(ground_position, ...[, ...])

Resample classification map in epipolar geometry

interpolate(points, values, positions[, method, ...])

Interpolate position
cars.applications.ground_truth_reprojection.ground_truth_reprojection_tools.get_ground_truth(geom_plugin, grid, sensor, disp_to_alt_ratio, target, window, geom_plugin_dem_median=None, reverse=False)[source]

Computes ground truth in epipolar and sensor geometry.

Parameters

  • geom_plugin_dem_median (geometry_plugin) – path to initial dem

  • geom_plugin (GeometryPlugin) – Geometry plugin with user’s DSM used to generate epipolar grids.

  • grid (CarsDataset) – Grid left.

  • sensor (dict) – sensor data Dict must contain keys: “image”, “color”, “geomodel”, “no_data”, “mask”. Paths must be absolute.

  • disp_to_alt_ratio (float) – Disp to altitude ratio used for performance map.

  • target (str) – sensor, epipolar or both outputs geometry

  • window (np.ndarray) – size of tile

  • geom_plugin_dem_median – Geometry plugin with dem median

  • reverse (bool) – true if right-> left

cars.applications.ground_truth_reprojection.ground_truth_reprojection_tools.resample_auxiliary_values(ground_position, auxiliary_input, window, interpolation_method='nearest', keep_band=False)[source]

Resample classification map in epipolar geometry

Parameters

  • ground_position (2D np.darray) – Direct localization result (lon, lat, alt)

  • auxiliary_input (string) – Path to auxiliary_value

  • window (list) – the tile window

  • interpolation_method (string) – interpolation method

  • keep_band (bool) – bool to see if we keep the band

cars.applications.ground_truth_reprojection.ground_truth_reprojection_tools.interpolate(points, values, positions, method='linear', fill_value=None)[source]

Interpolate position

Parameters

  • points (np.darray) – Points defining the grid

  • values (2D np.darray, or 3D np.darray (band, row, col)) – Data

  • positions (np.darray) – Positions to interpolate

  • method (string : {'linear', 'nearest'}) – interpolation method

  • fill_value (float) – value to use for points outside of the interp domain

Returns

interpolated positions

Return type

1D np.darray or 2D np.darray (band, interpolated positions)