pandora2d.cost_volume_confidence.ambiguity

This module contains functions associated to the cost volume confidence computation step with ambiguity method.

Classes

Ambiguity

Ambiguity class

Module Contents

class pandora2d.cost_volume_confidence.ambiguity.Ambiguity(cfg: dict)

Bases: pandora2d.cost_volume_confidence.cost_volume_confidence.CostVolumeConfidence

Ambiguity class

_normalization

_eta_min = 0.0

_eta_max

_eta_step

_percentile = 1

property schema

property defaults

confidence_prediction(left_image: xarray.Dataset, cost_volumes: xarray.Dataset, dataset_disp_maps: xarray.Dataset) → tuple[xarray.Dataset, xarray.Dataset]

Compute a confidence prediction.

Parameters:

• left_image – left Dataset image

• right_image – right Dataset image

• cost_volumes – cost volume dataset

• dataset_disp_maps – dataset containing row and col disparity maps

Returns:

the disparity map and the cost volume updated with the confidence measure

static normalize_with_extremum(confidence: numpy.ndarray, nbr_disparities: int, nbr_etas: int) → numpy.ndarray

Normalize ambiguity with extremum

Parameters:

• confidence – confidence

• nbr_disparities – number of disparity (row_disparity * col_disparity)

• nbr_etas – size of etas

Returns:

the normalized confidence