pandora.marge
This module contains the function which defines the images margins.
Module Contents
Functions
|
Calculates the margins for the left and right images according to the |
- pandora.marge.get_margins(disp_min: int, disp_max: int, cfg: Dict[str, dict]) xarray.Dataset [source]
Calculates the margins for the left and right images according to the steps present on the pipeline.
The margins are initialized at [+-disp_max, 0, -+disp_min, 0] for right and left images.
For matching_cost computation, a margin of half the window size is considered.
For vfit refinement, a margin of 1 pixel size is considered at positions 0 and 2.
For median filter, a margin of half the filter size is considered.
For SGM or 3SGM optimization, since infinite margins are theoretically required, a relatively high value (40 pixels) is considered.
The maximum margin obtained between (matching_cost + vfit refinement + median filter) and SGM/3SGM will be considered.
- Parameters:
disp_min (int) – minimal disparity
disp_max (int) – maximal disparity
cfg (dict of dict) – user configuration
- Returns:
margin for the images, 2D (image, corner) DataArray, with the dimensions image = [‘left_margin’, ‘right_margin’], corner = [‘left’, ‘up’, ‘right’, ‘down’]
- Return type:
xr.dataset