Gap Filling

Contents

Gap Filling#

Algorithms that replace the undefined values (NaN) of a grid, typically before interpolating near coastlines or data gaps. gauss_seidel(), multigrid() and fft_inpaint() solve the problem iteratively and modify the 2D array in place, loess() returns a new array, and matrix() and vector() fill by linear interpolation. The iterative methods and LOESS accept keyword arguments or a configuration object (see Configuration Objects).

Algorithms#

fft_inpaint(grid[, config, max_iterations, ...])

Fill missing values using FFT-based inpainting.

gauss_seidel(grid[, config, max_iterations, ...])

Fill missing values using Gauss-Seidel relaxation.

loess(data[, config, nx, ny, ...])

Fill missing values using LOESS (locally weighted regression).

multigrid(grid[, config, max_iterations, ...])

Fill missing values using multigrid method.

matrix(grid[, fill_value])

Fill a 2D array by linear interpolation.

vector(array[, fill_value])

Fill a 1D array by linear interpolation.
Contents