Speckle filtering is the spatial filtering of synthetic aperture radar imagery to suppress speckle, the grainy salt and pepper pattern produced by coherent imaging.12
Speckle is inherent to narrow-banded coherent imaging systems and results from interference among the many scattering echoes within a resolution cell.1 ESA’s STEP documentation describes it as random constructive and destructive interference of coherent return waves from the elementary scatterers in each cell.2 The SAR Handbook adds that speckle depends on image brightness, unlike most other noise sources, which is why it is difficult to treat as simple additive noise.1
Reduction can be applied either by spatial filtering or by multilook processing, and only the first is speckle filtering in the narrow sense.2 The STEP operator offers Boxcar, Median, Frost, Lee, Refined Lee, Gamma-MAP, Lee Sigma and IDAN filters.2 The SAR Handbook says that Lee and similar sigma filters, unlike Gaussian or boxcar filters, preserve image sharpness and detail while suppressing noise.1
None of the retained sources gives speckle filtering a NASA processing level, a preferred filter or a window size, and none offers a quantitative comparison of filters.
Footnotes
-
NASA Earthdata, “The SAR Handbook, Chapter 2, Spaceborne Synthetic Aperture Radar”. https://earthdata.nasa.gov/s3fs-public/2025-04/SARHB_CH2_Content.pdf ↩ ↩2 ↩3 ↩4
-
ESA STEP, “Speckle Filter Operator”. https://step.esa.int/main/wp-content/help/versions/9.0.0/snap-toolboxes/org.esa.s1tbx.s1tbx.op.sar.processing.ui/operators/SpeckleFilterOp.html ↩ ↩2 ↩3 ↩4