SAR Imaging
Synthetic Aperture Radar (SAR) imaging is an Earth Observation (EO) method, where a satellite sends radar pulses to the Earth and monitors it's reflections. The moving satellite forms so called synthetic aperture, which can be used for studying the signal reflections from different points on earth. The time of flight of signal provides a distance between satellite and the ground surface. Because the exact location of the satellite is known, the distance can be used for estimating Earth topography. On the other hand, the strength of the reflected signals carries information about the properties of the Earth surface.
There are many SAR satellites in the orbit, such as
- Two Sentinel 1 satellites from Copernicus program of the European Space Agency (ESA).
- ALOS PALSAR
- IceEye
- JERS
- TERRASAR-X
- TanDEM-X
Some of the SAR satellite data is available from public sources
- Data from all satellites from Copernicus program can be downloaded from SciHub, instructions for usage can be found from the Copernicus MOOC course.
- Data from many publicly open satellites can be accessed using Google Earth Engine, and it can be also used for making and running your own algorithms in the cloud. One example is the night light test application.
Some useful SAR resources
- ESA, free EO and SAR data:
- ALOS, JERS, TERRASAR-X/TandDEM-X
- https://earth.esa.int/web/guest/pi-community/apply-for-data
- Earth observation gateway
- Earth Online 3rd party data products
- The space portal of University of Vaasa
- NASA data sources, for example NASA Earth data
- Commercial ones, Planet, Airbus, ICEye etc. You can test and learn with the free data but it’s limited.
- University of Alaska's SAR satellites
SAR information
The wavelength of the sensor determines the penetration depth of the transmitted signal into the vegetation layer of the terrain surface. The longer the wavelength, the deeper the penetration can be, particularly in forests.
Frequency bands and some of their usage
- L-band: 1-2 GHz, Wavelength 15-30 m (GPS, 3G74G)
- S-Band: 2-4 GHz, 7.5-15 m (weather radars, 5G)
- C-band: 4-8 GHz, 3.75 - 7.5 m (Satellite communication, satellite television)
- X-band: 8-12 GHz, 2.5-3.75 m (SAR, military radars)
- Ku-band: 12-18 GHz 1.7 - 2.5 m (Satellite communication)
- Ka-band: 26-40 GHz, 0.75 - 1.15 m (Satellite communication)
The energy of an X-band sensor is mainly returned at the top layer of the canopies
Most of the L-band signal penetrates through the upper vegetation layer and is returned at the ground surface.
The backscatter behavior of C-band is less predictable. Due to volume scattering effects, the layer of backscattering is less determined and does not correspond directly to a terrain surface — neither the vegetation surface nor the ground surface.
The SAR satellites usually send the radio signal either vertically (V) or horizontally (H) polarized. They can also receive the reflected signal back using either vertical or horizontal polarizer. The polarization angles used in transmission and reception can therefore form the following four combinations:
- VV: the signal is send as vertically polarized and received as vertically polarized. This leads strong reflection from
- HH: The signal is send as horizontally polarized and received as horizontally polarized. This leads strong reception from smooth specular surfaces.
- VH: The signal is send as vertically polarized by the receiver only collects horizontally polarized signals. This cross-polarization combination mostly blocks the pure specular reflections which is keeping the polarization angle. On the contrary, the reflections from more complex surfaces having multiple scattering looses the polarization angle and is therefore partly received.
- HV: Similar cross polarization method than VH.
More in depth discussion of polarimetric SAR.
Sentinel 1
- Two satellites in polar orbits, 693 km height
- 12 days repeat cycle (each)
- The C-SAR’s advantage appears to be an ability to operate at C-band wavelength that cannot be blocked by cloud cover or insufficient light.
- 5.405 GHz, bandwidth 0-100 MHz, polarization HH+HV, VV+VH, VV, HH
- Azimuth beam width 0.23”
- Available data products with different level of preprocessing
- Level-0 Products: condensed source packets which have additional annotations and lists auxiliary information that supports processing. Includes manifest.safe file, measurement and index data components.
- Level-1 Products
- Slant-range single-look complex products (SLC): sdf phase an amplitude, I-Q values of each pixel.
- ground range Detected geo-referenced products (GRD): All data is projected to ground range using an Earth ellipsoid model corrected with specified terrain heigh
- Level-2 Products: geo-located geophysical products copied from Level-1. Level-2 Ocean (OCN) products applicable for wind, wave and currents can have geophysical components obtained from the SAR data like: Ocean Wind field (OWI), Ocean Swell spectral range (OSW), Surface Radial Velocity (RVL), Availability of each one highly depends on mode of acquisition
Tutorials and learning resources
- ESA tutorials
- What is SAR?, material from the University of Alaska
- Tutorial for downloading data from SentinelHUB
- How to get free research access to Copernicus using Network of Resources (NOR)
- Custom scripts for ice evaluation using S1 SAR data
Google tools for accessing and processing EO data
- Google Earth Engine online notebooks and Python and JavasScript APIs
ESA tools for accessing and processing EO data (These are installed in the RUS environment)
- Access Sentinel Hub from QGIS, also available in Puhti in CSC for those who have access to CSC services.
- Python Sentinel API in Jupyter notebook (example)
- Python Sentinel API for downloading Sentinel data by Sentinel HUB
- Python and JavaScript Sentinel API:s
- SNAP, also available in Puhti in CSC for those who have access to CSC services
- BRAT (Basic Radar Altimetry Toolbox)
- Libraries:
- GDAL (raster and vector geospatial data library), also available in Puhti in CSC for those who have access to CSC services
- Orfeo Toolbox (OTB) (Open Source processing of remote sensing images) also available in CSC
- Sen2Cor, SNAP plugin, processor for Sentinel-2 Level 2A product generation and formatting, also available in Puhti in CSC for those who have access to CSC services
- SNAPHU, SNAP plugin, two-dimensional phase unwrapping to recover unambiguous phase data from a 2-D array of phase values known only modulo 2pi rad.
RUS environment can be used as a virtual server in ESA:s server room or as a Docker image. Both are Linux environments, with all above tools and many others already installed. RUS is only available for EU citizens. The tools can be also installed in normal Linux computer, and then there is no need for Docker or other virtualization platform, and the resources of the computer can therefore used more efficiently.