Analysis-ready Sentinel-1 GRD data generation

Available with Image Analyst license.

A Sentinel-1 Level 1 synthetic aperture radar (SAR) image must be processed before it can be used for visualization or analysis. Some problems that need to be addressed include removing thermal noise, calibrating to retrieve a meaningful backscatter value, removing speckle noise, removing radiometric and geometric distortions, and rendering images with a large value range.

The Synthetic Aperture Radar toolset, in the Image Analyst toolbox, contains eight tools to generate calibrated, terrain-corrected, analysis-ready imagery data from Sentinel-1 Ground Range Detected (GRD) data. The following tools are used to generate analysis-ready Sentinel-1 GRD data, as shown in the diagram below:

  • Download Orbit File
  • Apply Orbit Correction
  • Remove Thermal Noise
  • Apply Radiometric Calibration
  • Apply Radiometric Terrain Flattening
  • Despeckle
  • Apply Geometric Terrain Correction
  • Convert SAR Units

Analysis-ready workflow

Orbit state vector download and use

The accuracy of radiometric and geometric terrain corrections relies on the supplied orbit state vectors (OSVs). Three types of OSVs are available for the Sentinel-1 product: predicted, restituted, and precise. Predicted OSVs are provided with the Sentinel-1 Level 1 GRD and SLC auxiliary products, restituted OSVs are available through the European Space Agency (ESA) within three hours of image acquisition, and precise OSVs are available through ESA within three weeks of image acquisition. It is recommended that you update the OSVs to restituted or precise once they are available.

The Download Orbit File tool identifies and downloads the appropriate OSV file. The Apply Orbit Correction tool uses this downloaded OSV file to update the Sentinel-1 product metadata.

Thermal noise removal

SAR images are distorted by additive thermal noise. Thermal noise is most apparent in images with low backscatter distribution, such as in the cross-polarized channel, which is characterized by a narrower backscatter distribution.

Due to the Sentinel-1 Terrain Observation with Progressive Scans (TOPS) acquisition mode, the thermal noise in Sentinel-1 products varies for individual subswath scans. Thermal noise commonly manifests as a sharp contrast between the subswath scans. The Remove Thermal Noise tool uses Sentinel-1 product metadata to correct thermal noise.

Radiometric calibration

The Apply Radiometric Calibration tool uses the Sentinel-1 product metadata to retrieve meaningful backscatter values. Radiometric calibration is the process of converting SAR products from image pixel digital number (DN) to the physical quantity of SAR backscatter intensity per unit area. The three calibration types are beta nought (Beta nought), sigma nought (Sigma nought), and gamma nought (Gamma nought). The unit area used for the calibration determines the calibration type.

Beta nought represents the radar reflectivity per unit area in slant range and is commonly known as radar brightness coefficient.

Sigma nought represents the radar reflectivity per unit area in ground range. Although sigma nought is a popular option for describing reflectivity, use it with caution. Sigma nought values vary with incidence angle, so a feature in the near range may have a different sigma nought value in the far range. If performing multitemporal analysis or change detection using sigma nought, use images from the same sensor and the same viewing geometry to ensure that changes in sigma nought are due to physical processes over time and not artifacts resulting from differences in viewing geometry.

Gamma nought represents the radar reflectivity per unit area in the plane perpendicular to the slant range. It is normalized using the incidence angle relative to the ellipsoid, so it provides a measurement value that is independent of range. If you want to use backscatter values to distinguish between unique features in a single image, use gamma nought instead of sigma nought. Also, use gamma nought if you are interested in multitemporal analysis or change detection using SAR imagery from different sensors or different viewing geometries (ascending versus descending). Gamma nought should only be used in these types of applications if the terrain is flat.

Radiometric terrain flattening

Due to the side-looking nature of SAR sensors, features facing the sensor appear artificially brighter than features facing away from the sensor. The Apply Radiometric Terrain Flattening tool corrects artificial radiometric values originating from complex topography and the sensor's viewing geometry.

Given an input DEM and an input Sentinel-1 GRD product calibrated to beta nought, the Apply Radiometric Terrain Flattening tool uses the range-Doppler approach1 to compute the illuminated area to produce a terrain-flattened gamma nought output. Alternatively, you can specify a terrain-flattened sigma nought output, which is normalized using the DEM-based local incidence angle.

An optional output is the simulated scattering area. These outputs can be used to gain insight on how terrain artificially impacts the nonterrain-flattened calibrated data.

Another optional output is a geometric distortion mask for identifying pixels affected by shadow, foreshortening, lengthening, or layover. The geometric distortion mask output allows you to mask terrain-flattened gamma nought or sigma nought output based on the geometric distortion type.

The last optional output is a geometric distortion raster containing a proxy for terrain slope, the look angle, the foreshortening ratio, and the local incidence angle. The geometric distortion output provides data that is used to perform the terrain flattening and to identify pixels impacted by geometric distortions.

Radiometric terrain flattening must be performed for applications interpreting a single image over any terrain, or for applications comparing multiple images from different sensors or the same sensor with different viewing geometries over any terrain.


SAR images are characterized by noisy anomalies known as speckle. This inherent condition results from the constructive and destructive interference of the backscattered signal. The Despeckle tool provides several speckle filters to improve the signal-to-noise ratio of the SAR image. The speckle filters available are Lee2, Enhanced Lee3, Refined Lee4, Frost5, Kuan6, and Gamma MAP7. These filters depend on local pixel statistics to optimize the speckle suppression while conserving feature detail. To preserve the statistical properties needed for these filters, it is recommended that you apply despeckling before geometric terrain correction, which resamples and reprojects the data.

Geometric terrain correction

Since SAR sensors are side looking, features facing the sensor appear compressed, while features facing away from the sensor appear stretched. The Apply Geometric Terrain Correction tool corrects geometric distortions, shifting the pixels to their correct geolocation.

The Apply Geometric Terrain Correctiontool uses the range-Doppler approach and the input DEM to orthorectify the input SAR image. A DEM with a resolution close to, or higher than, the input SAR data is recommended for most applications. For applications in which no terrain is present, you can omit the input DEM. The Apply Geometric Terrain Correction tool can use the range-Doppler approach and the geolocation grid from the product metadata to orthorectify the input SAR image.

Conversion to decibels

The final step to prepare analysis-ready data is to convert the unitless (linear) backscatter intensity to decibels (dB). The Convert SAR Units tool converts the linear backscatter intensity to decibels using a simple logarithmic conversion. The logarithmic conversion reduces the range of backscatter intensity values to improve image visualization and interpretation.


[1] Small, D. 2011. "Flattening Gamma: Radiometric Terrain Correction for SAR Imagery." IEEE Transactions on Geoscience and Remote Sensing 49, no. 8: 3081–3093. DOI: 10.1109/TGRS.2011.2120616.

[2] J. S. Lee. "Digital Image Enhancement and Noise Filtering by Use of Local Statistics." IEEE Transactions on Pattern Analysis and Machine Intelligence vol. PAMI-2, no. 2, pp. 165–168, March 1980, DOI: 10.1109/TPAMI.1980.4766994.

[3] A. Lopes, R. Touzi, and E. Nezry. "Adaptive speckle filters and scene heterogeneity." IEEE Transactions on Geoscience and Remote Sensing vol. 28, no. 6, pp. 992–1000, Nov. 1990, DOI: 10.1109/36.62623.

[4] J. S. Lee and E. Pottier. "Polarimetric radar imaging: from basics to applications." CRC press, Dec. 2017.

[5] V. S. Frost, J. A. Stiles, K. S. Shanmugan, and J. C. Holtzman. "A Model for Radar Images and Its Application to Adaptive Digital Filtering of Multiplicative Noise." IEEE Transactions on Pattern Analysis and Machine Intelligence vol. PAMI-4, no. 2, pp. 157–166, March 1982, DOI: 10.1109/TPAMI.1982.4767223.

[6] D. T. Kuan, A. A. Sawchuk, T. C. Strand, and P. Chavel. "Adaptive Noise Smoothing Filter for Images with Signal-Dependent Noise." IEEE Transactions on Pattern Analysis and Machine Intelligence vol. PAMI-7, no. 2, pp. 165–177, March 1985, DOI: 10.1109/TPAMI.1985.4767641.

[7] A. Lopes, E. Nezry, R. Touzi, and H. Laur. "Maximum A Posteriori Speckle Filtering And First Order Texture Models In Sar Images." 10th Annual International Symposium on Geoscience and Remote Sensing, 1990, pp. 2409–2412, DOI: 10.1109/IGARSS.1990.689026.

Related topics