Raster Gap Filler
Upload a raster with two or more bands — a satellite band with cloud gaps, a field survey rasterized with mostly-empty cells, anything with real holes in it — pick the band to fill (the target) and, if you have one, a second band with no gaps that’s correlated with it (the covariate, elevation or another spectral band are common choices). This tool fits a variogram model between the two, then predicts the target’s missing values from wherever the covariate is available — cokriging, not just averaging nearby pixels. Skip the covariate and it falls back to ordinary kriging on the target band’s own known pixels.
Private by design. Your file is processed entirely in your browser (R via WebAssembly) — it’s never uploaded to any server, including ours. Nothing to configure, nothing to trust: the code never gets the chance to see your data.
The code itself isn’t shown here — it’s a fixed, non-editable computation using gstat’s cokriging (linear model of coregionalization) underneath. Want to see the same pipeline built up one step at a time, why a covariate band helps, and why this runs on R, not Python? Read Filling Raster Gaps with Cokriging.
First run loads the R engine — expect 1-3 minutes (R’s WebAssembly build is noticeably slower to start than this site’s Python tools). Instant after that, and again each time you come back to this page.
Got sparse point samples instead of a raster with gaps? The Kriging Interpolator predicts a full surface from scattered measurements. Want to look at a raster’s bands before deciding what to fill? Start with the Raster Inspector.