Fill NoData

Draft: written, not yet checked by a person Operation · fill-nodata

Fills nodata cells of a raster by interpolating from the valid cells around them.

Pitfall. Engines interpolate differently (inverse distance from the gap's edge, splines, bilinear, nearest), so large gaps come out differently.

Pitfall. Most engines stop at a maximum distance and leave larger gaps as nodata.

About this operation
Also called	fill gaps, fill missing data, interpolate nodata, fill nulls, Elevation Void Fill
Group	`raster/surface`
Checks	`raster.nodata`
Pitfalls	2 listed above
Columbus tool	Not stated
Notes	Not stated
Status	active
Broader	Not stated
Narrower	Not stated
Related	`idw-interpolation` `focal-statistics`

Parameters

Not stated

Direction	Name	Type	Geometry	Note
Input	`raster`	raster	Not stated	Not stated
Output	`output`	raster	Not stated	Not stated

Engine	Tool	Parameters	Differences
`gdal`	`gdal raster fill-nodata`	`raster` → `input` `output` → `output`	A single missing cell on a plane (neighbours 9.5 and 11.5 west and east, 10.5 north and south) was filled with 10.96, not 10.5 (GDAL 3.13). Inverse distance weighting from the gap's edge (strategy invdist) or nearest; max-distance (100 cells) limits the search.
`grass`	`r.fillnulls`	`raster` → `input` `output` → `output`	Regularized spline with tension by default (method rst); bilinear and bicubic are also offered.
`qgis`	`gdal:fillnodata`	`raster` → `INPUT` `output` → `OUTPUT`	A single missing cell on a plane (neighbours averaging 10.5) was filled with 10.96, as GDAL's fill does. Runs GDAL's fill nodata; DISTANCE (10 cells by default) limits the search.
`rasterio`	`rasterio.fill.fillnodata`	`raster` → `image`	GDAL's algorithm on a NumPy array; mask marks the valid cells (0 for the cells to fill). A single missing cell on a plane (neighbours averaging 10.5) was filled with 10.96, as GDAL's fill does.
`rioxarray`	`rioxarray.raster_array.RasterArray.interpolate_na`	Not stated	A method of the DataArray's rio accessor; method defaults to nearest, so a gap takes one neighbour's value; linear and cubic (scipy's griddata) interpolate.
`whiteboxtools`	`FillMissingData`	`raster` → `input` `output` → `output`	Inverse distance weighting within filter (11 cells); no_edges (on by default) leaves gaps that touch the raster's edge.