cars.applications.grid_generation.grid_correction

Grids module: contains functions used for epipolar grid correction

Module Contents

Functions

correct_grid_from_1d(grid, grid_correction_coef[, ...])

Correct grid from correction given in 1d

correct_grid(grid, grid_correction[, save_grid, ...])

Correct grid

estimate_right_grid_correction(matches, grid_right[, ...])

Estimates grid correction, and correct matches

create_matches_cars_ds(corrected_matches, initial_cars_ds)

Create CarsDataset representing matches, from numpy matches.
cars.applications.grid_generation.grid_correction.correct_grid_from_1d(grid, grid_correction_coef, save_grid=False, pair_folder=None)

Correct grid from correction given in 1d

Parameters

  • grid (CarsDataset) – grid to correct

  • grid_correction_coef (list(float), size 6) – grid correction to apply

  • save_grid (bool) – if True grids are saved in root or pair_folder

  • pair_folder (str) – directory where grids are saved

cars.applications.grid_generation.grid_correction.correct_grid(grid, grid_correction, save_grid=None, pair_folder=None)

Correct grid

Parameters

  • grid (CarsDataset) – grid to correct

  • grid_correction (Tuple(np.ndarray, np.ndarray) (coefsx_2d, coefsy_2d) , each of size (2,2)) – grid correction to apply

  • save_grid (bool) – if True grids are saved in root or pair_folder

  • pair_folder (str) – directory where grids are saved

cars.applications.grid_generation.grid_correction.estimate_right_grid_correction(matches, grid_right, initial_cars_ds=None, save_matches=False, pair_folder='', pair_key='pair_0', orchestrator=None)

Estimates grid correction, and correct matches

Parameters

  • matches (np.ndarray) – matches

  • grid_right (CarsDataset) – grid to correct

  • save_matches (bool) – true is matches needs to be saved

  • pair_folder (str) – folder used for current pair

Returns

grid_correction to apply, corrected_matches, info before, info after

Return type

Tuple(np.ndarray, np.ndarray) , np.ndarray, dict, dict grid_correction is : (coefsx_2d, coefsy_2d) , each of size (2,2)

cars.applications.grid_generation.grid_correction.create_matches_cars_ds(corrected_matches, initial_cars_ds)

Create CarsDataset representing matches, from numpy matches. Matches are split into tiles, stored in pandas DataFrames

Right CarsDataset is filled with Nones

Parameters

  • corrected_matches (numpy array) – matches

  • initial_cars_ds (CarsDataset) – cars dataset to use tiling from

:return new_matches_cars_ds :rtype: CarsDataset