pandora2d.matching_cost.base

Module for common base of all MatchingCost methods.

Classes

BaseMatchingCost

MatchingCost base class.

Functions

`set_out_of_row_disparity_range_to_other_value`(→ None)	Put special value in data where the row disparity is out of the range defined by disparity grids.
`set_out_of_col_disparity_range_to_other_value`(→ None)	Put special value in data where the column disparity is out of the range defined

Module Contents

class pandora2d.matching_cost.base.BaseMatchingCost(cfg: dict)[source]

Bases: abc.ABC

MatchingCost base class.

_cfg[source]

_method[source]

_step_row[source]

_step_col[source]

_window_size[source]

_subpix[source]

_spline_order[source]

_float_precision[source]

cost_volumes: xarray.Dataset | None = None[source]

shifted_right_images: list[xarray.Dataset] = [][source]

property schema[source]

property defaults[source]

check_conf(cfg: dict) → dict[str, str][source]

Check the matching cost configuration

Parameters:: cfg – user_config for matching cost
Returns:: cfg: global configuration

_update_with_default_config_values(cfg: dict)[source]

property cfg: collections.abc.Mapping[str, str | int | list[int]][source]

Get used configuration

Returns:: cfg: dictionary with all parameters

property step: list[int][source]

Get step [row, col]

Returns:: step: list with row & col step

property window_size: int[source]

Get window_size

Returns:: window_size: window used to compute correlation

allocate_cost_volumes(cost_volume_attr: dict, row: numpy.ndarray, col: numpy.ndarray, disp_range_row: numpy.ndarray, disp_range_col: numpy.ndarray, np_data: numpy.ndarray = None) → xarray.Dataset[source]

Allocate the cost volumes

Parameters:

cost_volume_attr – the cost_volume’s attributes
row – dimension of the image (row)
col – dimension of the image (columns)
disp_range_row – rows disparity range.
disp_range_col – columns disparity range.
np_data – 4D numpy.ndarray og cost_volumes. Defaults to None.

Returns:

cost_volumes: 4D Dataset containing the cost_volumes

get_cv_row_col_coords(img_row_coordinates: numpy.typing.NDArray, img_col_coordinates: numpy.typing.NDArray, cfg: dict) → tuple[numpy.typing.NDArray, numpy.typing.NDArray][source]

Compute cost_volumes row and col coordinates according to image coordinates

Parameters:

img_row_coordinates – row coordinates of left image
img_col_coordinates – col coordinates of left image
cfg – matching_cost computation configuration

Returns:

a Tuple of np.ndarray that contains the right coordinates for row and col

get_disp_row_coords(img_left: xarray.Dataset, margins: pandora2d.margins.Margins) → numpy.typing.NDArray[source]

Compute cost_volumes row disparity coordinates according to image disparities

Parameters:

img_left – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
margins – refinement margins

Returns:

a Tuple of np.ndarray that contains the right coordinates for disparities

get_disp_col_coords(img_left: xarray.Dataset, margins: pandora2d.margins.Margins) → numpy.typing.NDArray[source]

Compute cost_volumes col disparity coordinates according to image disparities

Parameters:

img_left – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
margins – refinement margins

Returns:

a Tuple of np.ndarray that contains the right coordinates for disparities

static cfg_for_get_coordinates(cfg: dict) → dict[source]

Return right configuration to give to get_coordinates or get_coordinates_2d methods.

To get right coordinates in cost_volume when initial left_margin > cfg[“ROI”][“col”][“first”] or initial up_margin > cfg[“ROI”][“row”][“first”] We need to have left_margin = cfg[“ROI”][“col”][“first”] and up_margin = cfg[“ROI”][“row”][“first”]

Parameters:: cfg – user configuration
Returns:: updated configuration to be sure to process the first point of ROI when ROI margin > ROI first point (left or up)

abstract property margins[source]

allocate(img_left: xarray.Dataset, img_right: xarray.Dataset, cfg: dict, margins: pandora2d.margins.Margins = None) → None[source]

Allocate the cost volume

Parameters:

img_left – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
img_right – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
cfg – matching_cost computation configuration
margins – refinement margins

Returns:

None

set_out_of_disparity_range_to_other_value(img_left: xarray.Dataset, value: int | float) → None[source]

Put special value in data where the row or column disparity is out of the range defined by disparity grids.

The operation is done inplace.

Parameters:

img_left – left image xarray.Dataset
value – value to set on data.

abstract set_shifted_right_images(img_right: xarray.Dataset) → None[source]

Compute shifted by subpix right image and assign shifted_right_images attribute.

Parameters:: img_right – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
Returns:: None

abstract compute_cost_volumes(img_left: xarray.Dataset, img_right: xarray.Dataset, margins: pandora2d.margins.Margins = None) → xarray.Dataset[source]

Computes the cost volumes

Parameters:

img_left – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
img_right – xarray.Dataset containing : - im : 2D (row, col) xarray.DataArray - msk : 2D (row, col) xarray.DataArray
margins – refinement margins

Returns:

cost_volumes: 4D Dataset containing the cost_volumes

pandora2d.matching_cost.base.set_out_of_row_disparity_range_to_other_value(data: xarray.DataArray, min_disp_grid: numpy.typing.NDArray[numpy.floating], max_disp_grid: numpy.typing.NDArray[numpy.floating], value: int | float, global_disparity_range: list[int]) → None[source]

Put special value in data where the row disparity is out of the range defined by disparity grids.

The operation is done inplace.

Parameters:

data – cost_volumes to modify.
min_disp_grid – grid of min disparity.
max_disp_grid – grid of max disparity.
value – value to set on data.
global_disparity_range – global row disparity range

pandora2d.matching_cost.base.set_out_of_col_disparity_range_to_other_value(data: xarray.DataArray, min_disp_grid: numpy.typing.NDArray[numpy.floating], max_disp_grid: numpy.typing.NDArray[numpy.floating], value: int | float, global_disparity_range: list[int]) → None[source]

Put special value in data where the column disparity is out of the range defined by disparity grids.

The operation is done inplace.

Parameters:

data – cost_volumes to modify.
min_disp_grid – grid of min disparity.
max_disp_grid – grid of max disparity.
value – value to set on data.
global_disparity_range – global column disparity range