Processing Sentinel-1 SAR images using Sentinel Application Platform (SNAP)

In this tutorial, I will walk through some of the basic steps needed to process the Sentinel-1 SAR data using SNAP. For this tutorial, I am using SNAP version 7.0. You can check the version of the SNAP from Help -> About SNAP…

1. If you do not have already, go ahead and download the SNAP from the here. You will need Java Runtime Environment for running SNAP.
2. There are many sources for downloading SAR imagery. For this tutorial, I will be focusing on using Sentinel-1, which is free, has high resolution enough to serve my need. We will be using the Alaska Satellite Facility (ASF) in order to download the Sentinel-1 datasets. For that, we need NASA Earth Data login credentials. So, go ahead, and create an account on the NASA Earth Data Login.
3. Next, go ahead and open up the ASF Data Search Platform. Click the User icon as shown in the red box below, and log in with the Earth Data credentials that you created in 1.
4. Next, select your area of interest (AOI), select Sentinel-1 images, and the dates you want the data for. Feel free to change any other parameters like Orbital Pass, Polarization, etc. Hit SEARCH and it will show up the list of the images. I will use 3 ascending images between 2019-01-01 to 2019-01-30, for my AOI.
5. After you have finished downloading, go ahead and open up SNAP. You may see that the downloaded data are in zip format. You do not need to unzip them, as SNAP can take in the zipped datasets. Go to File -> Open Product. Click Open to import them to the SNAP.
6. Next step is to apply the orbit file. The orbit file contains information on the satellite orbit information. An accurate orbit file helps in better geocoding and hence the processing of the SAR dataset. Go to Radar -> Apply Orbit File. The good thing is SNAP can auto-download the latest orbital file. Make sure you have correct pathname for saving the orbital applied file. Refer to the image below. Hit Run.
   Apply to all the scenes that you downloaded. In the Product Explorer on the left side, you can see the new file with _Orb indicating the orbital applied file if you choose to use the default file name.
7. After this, we need to calibrate our orbitally corrected data. Go to Radar -> Radiometric -> Calibrate to access the calibration menu. Choose to output to beta0 band and select all the polarization that is available. We will calibrate the beta0 band to gamma0 image during the terrain flattening step. Apply to all the orbital corrected files. In the Product Explorer on the left side, you can see the new file with _Cal indicating the calibrated file if you choose to use the default file name.
8. Next step is the multilooking step. Go to Radar -> SAR Utilities -> Multilooking to access the multilooking menu. It helps to remove the speckle noise in the image. Select all the source band that you have and change the number of range and azimuth looks to 3. Feel free to experiment with different looks to see which gives the best result.
9. The next step is to perform the Terrain Flattening. The terrain effect must be resolved before we can perform any processing. Usually, a high-resolution DEM with some resampling method is used to perform the terrain flattening. To perform terrain flattening, go to Radar -> Radiometric -> Radiometric Terrain Flattening. As mentioned in 7, we are calibrating the image to gamma0 using 1 arcsec DEM from SRTM.
10. In the Range-doppler projection, the imagery from the ascending and descending orbits are flipped vertically with respect to each other. The technique is to find how the corresponding pixel shift between the images that may be caused by the topography. Hence we will be using the DEM assisted coregistration. This is also known as Back Geocoding. Go to Radar -> Coregistration -> DEM-Assisted Coregistration -> DEM Assisted Coregistration.
11. The next step is to speckle filter. Speckle is very common in SAR images because of the interference phenomenon when returning back from the surface imagined. This result in a salt-and-pepper effect that makes it difficult for visual interpretation as well as affect the resolution of the image. The speckle noise can be reduced by applying spatial filtering or multilook processing. For this tutorial, we will be using a spatial filter specifically the Lee-Sigma filtering. Go to Radar -> Speckle Filtering -> Single Product Speckle Filter.
12. The last step is to orthorectify the images. Go to Radar -> Geometric -> Ellipsoid Correction -> Geolocation-Grid. The goal is to generate nadir image so measurements can be done from images, and the effect of topography as well as hill are removed.
13. Now we can go ahead and export the images that we just processed. Go to File -> Export -> ENVI (or GeoTiff). You can now use the exported image in your favorite software like ENVI, QGIS, or ArcGIS, etc.