Fill NoData pixels with inverse distance weighted interpolation
Inverse distance weighting algorithm
Files processed in browser
Estimates values from surrounding pixels
Upload a raster with NoData areas
Set search distance and smoothing
Maximum radius to search for valid pixels
Number of mean smoothing passes on filled result
Fill NoData closes the gaps in a raster by interpolating missing pixels from their neighbours using inverse distance weighting (IDW). You control the maximum search distance in pixels and a smoothing pass from 0 to 5; because every calculation happens in the browser, sensitive elevation or survey data never leaves your computer.
It repairs the classic artifacts of raster workflows: seams and nodata strips left behind after clipping or mosaicking, holes along the edges of satellite scenes, and blank cells that break contouring, hillshading, or hydrological modelling.
Upload the raster with gaps - its nodata value is read automatically from the file's metadata, so no manual configuration is needed.
Set the maximum distance: only nodata pixels within this many pixels of valid data are interpolated, which stops huge voids from being smoothed unrealistically; optionally add 0-5 smoothing iterations to blend the transitions.
Run the fill and download the gap-free raster, computed entirely on your machine.
AIt is a search radius in pixels. A nodata pixel farther than this from any valid pixel stays empty, so a small distance fills only thin seams while a large distance can span wide voids at the price of heavier estimation. Match it to the actual width of the gaps in your data.
AInverse distance weighting averages the surrounding valid pixels, giving nearer pixels more influence. The result is a smooth gradient rather than a flat constant, which suits continuous data like elevation and temperature - class maps should never be interpolated this way.
AThose nodata areas lay beyond the maximum distance from any valid pixel. Raise the distance and run again; for rasters with both hairline seams and large voids, two passes with increasing distances often give the cleanest result.