GUIDE: FUNDAMENTALS OF SYNTHETIC APERTURE RADAR (SAR)

For anyone who wants to understand the complex concepts of SAR in a not so complex way

INTRODUCTION

Synthetic aperture radar (SAR) is a type of radar sensor that actively sends electromagnetic waves to the earth's surface and receives the reflected signal. The electromagnetic wave received by the sensor is called the measured backscatter. A SAR image is a 2D rendering of the measured backscatter.

An optical sensor (left) only works with clear, sunny skies. A SAR sensor (right) can work day and night, and even when clouds are present.

A sensor can be classified as either passive or active. A passive sensor, using optical systems, records electromagnetic waves emitted by the sun and reflected from the ground surface. An active sensor, used by SAR systems, functions as both the source and the receiver. This means the sensor transmits the electromagnetic waves and also records the reflected waves. Unlike an optical sensor, a SAR sensor can operate during the day or night, independent of the sun, since it transmits its own signal.

Powered by Esri
Powered by Esri

Example of an LANDSAT 8 optical image (left) vs Sentinel-1SAR image (right) of Balikpapan Bay, Indonesia

APPLICATIONS

DATA SOURCE:  Copernicus  Sentinel data [2017-2021]. Retrieved from  ASF DAAC  [20 December 2022], processed by esri.  ICEYE  data [2021-2022]. Retrieved from  ICEYE  [15 December 2022], processed by esri.

1

FLOODING

In August 2017, Hurricane Harvey hit the coast of Texas causing catastrophic flooding. The map shows the spatial extent of the flooding of land/permanent water bodies (blue color) and kicked up debris (pink color) before (August 5th, 2017) and after (August 29th, 2017) the hurricane.

2

OIL SPILLS

On August 23rd, 2021, an oil spill occurred from the Baniyas Thermal Station, a major oil refinery in Baniyas, Syria. The map shows the spatial extent of the oil spill (black and magenta color) on August 24th, 2021 and on August 30th, 2021.

3

DEFORESTATION

In Acre, Brazil, within the Amazon rainforest, there is deforestation mapped clearly by SAR data in just an 11 day period. Deforestation between July 22nd and August 2nd, 2021 is indicated by the black and magenta colors.

4

SEA ICE

Between December 17th, 2021 and February 13th, 2022 the sea ice in the Hudson Bay is starting to melt and break up into small pieces. SAR imaging of sea ice during dark arctic winters provides useful navigation insight.

5

SHIP DETECTION

Ships waiting to sail through the Panama Canal can be seen as yellow sprinkles just southeast of the entrance. SAR capabilities allow ship detection through storms when cloudy conditions are common.

ELECTROMAGNETIC SPECTRUM

Through active sensing, a SAR sensor can collect imagery using longer wavelengths compared to an optical sensor.

Optical sensor wavelengths = from visible to thermal infrared waves SAR sensor wavelengths = microwaves from K-band to P-band

Electromagnetic spectrum showing short to long wavelengths and the respective types of waves (from left to right – gamma rays, x-rays, ultraviolet, visible, infrared, microwave, radio) as well as the different bands that SAR sensors can use are available within the microwave range.

BAND WAVELENGTHS (cm)

Range of wavelengths for microwave bands categorized into short, medium and long wavelengths

The microwave wavelengths characterizing SAR systems also provide distinct information about the physical properties of the earth's surface, such as:

  • roughness
  • density
  • moisture content

Why?

Microwave wavelengths typically scatter differently based on the feature that reflects them. If the wavelength is longer than the feature of interest, the feature will be undetected by the electromagnetic wave.

The figure shows the shorter wavelengths having a weakened backscatter than the medium to long wavelengths which can easily penetrate clouds, fog, dust, smog, and smoke, making it better suited for monitoring humid tropics and high latitudes.

The figure shows how short, medium and long wavelengths have different penetration depths on dry alluvium (dirt). The longer the wavelength, the greater the penetration depth is. However, the greater the moisture content of the material, the shallower the penetration is. This characteristic is used to differentiate frozen and unfrozen ground conditions.

The figure shows the different bands interacting with a forest canopy. The band with a short wavelength reflects off the canopy, the band with a medium wavelength reflects within the canopy, and the band with a long wavelength penetrates through the canopy and to the forest floor.

Another important aspect is the interaction of the signal to an electric field. Objects on the ground react to an electric field in different ways. This can be attributed to the dielectric constant — how strongly a medium will reflect an electric field. So, the higher the dielectric constant, the stronger the reflection.

DIELECTRIC CONSTANTS

Vacuum/Air ≈ 1

Dry Soil ≈ 5

Wet soil ≈ 25

Water/metal ≈ 80

BACKSCATTER

The electromagnetic wave received by the sensor is called the measured backscatter.

For the amplitude of a SAR image pixel:

High digital number = strong backscatter Low digital number = weak backscatter

The strength of the amplitude of the measured backscatter is used to differentiate between features on the ground. The time delay between the transmitted and received electromagnetic wave determines the location of the feature.

A SAR image is commonly delivered as two product types:

  1. Ground Range Detected (GRD) - a GRD image is stored as a real valued array in which the value in each pixel represents the amplitude of the measured backscatter signal
  2. Single Look Complex (SLC) - an SLC image is stored as a complex valued array in which the single complex value in each pixel represents the amplitude and phase of the measured backscatter signal

GRD products have been averaged to produce a multilooked image projected to the ground range using an earth ellipsoid model. SLC products are in the image plane of the data acquisition, known as the slant range plane.

*note GRD images are the primary focus of this StoryMap

POLARIZATION

Active sensing also provides the ability to control the polarization of the transmitted electromagnetic waves. By having the SAR sensor define both the transmitted and received polarization, the resulting SAR image can highlight different features on the earth's surface based on the backscatter.

SAR data polarization is denoted by two letters:

First letter = transmitted polarization Second letter = received polarization

Animation showing the sensor transmitting vertically polarized waves and receiving vertically polarized (VV) or horizontally polarized (VH) waves, as well as transmitting horizontally polarized waves and receiving horizontally polarized waves (HH) or vertically polarized waves (HV).

Co-polarized - the transmitted and received waves share the same polarization:

VV polarized data HH polarized data

Cross-polarized - the transmitted and received waves do not share the same polarization:

VH polarized data HV polarized data

Dual-polarized SAR images feature either:

VV, VH polarized data HH, HV polarized data

Quad-polarized SAR images feature both:

VV, VH polarized data HH, HV polarized data

As with the wavelength, the transmitted and received polarization used strongly impacts the features captured in the SAR image and must be taken into consideration.

TYPES OF SCATTERING

For most microwave wavelengths, smooth, horizontal features such as roads, airport runways, dry lake beds, flattened soil, still water, and sand reflect the electromagnetic waves away from the sensor and exhibit pixels with weak backscatter. Similarly, for most microwave wavelengths, human-made objects characterized by reflective material and sharp geometries, such as buildings and ships, reflect the electromagnetic waves back to the sensor and exhibit pixels with strong backscatter.

There are FOUR main types of scattering:

  1. single-bounce
  2. double-bounce
  3. diffuse
  4. volume

Double-Bounce Scattering - the radar signal reflects once off a vertical target onto a smooth surface and reflects a second time off the smooth surface back toward the sensor. For Double-Bounce Scattering, the polarization is not altered and causes a high backscatter in the co-polarized band and a low backscatter in the cross-polarized band.

Diffuse Scattering - the radar signal reflects once off a rough surface and is scattered in various directions. The rougher the surface, the higher the co-polarized backscatter.

Diffuse scattering (lower left) can occur on rough surfaces such a dirt and gravel. Volume Scattering - the radar signal reflects off a 3D feature multiple times, changing polarization randomly during the reflection.

Volume scattering (upper right) can take place within canopies of short or sparse vegetation such as bushes, shrubs, or agricultural crops.

Single-Bounce Scattering - the radar signal reflects once off a smooth surface, reflecting in a perpendicular direction relative to the original signal. Single-Bounce Scattering is also known as "Specular Reflection" or "Forward Scattering" and causes a low backscatter in the co-polarized band and a low backscatter in the cross-polarized band.

SATELLITE GEOMETRY

SAR sensors send signals (radar rays) to illuminate the ground.

RADAR IMAGING ANGLES

  • Azimuth Direction - the direction along the flight path (satellite orbit)
  • Frame - collection of swaths
  • Ground Range - range along the ground surface
  • Incidence Angle - angle between the line of sight ray and the local vertical
  • Line of Sight (LOS) - look direction from the satellite to the ground
  • Local Vertical - vertical direction from the measured point on the ground
  • Look Angle - angle between the nadir direction and the radar ray
  • Nadir - vertical direction below the sensor
  • Path - flight direction containing image frames
  • Slant Range - illumination to the ground
  • Swath - window of reflected rays
  • Range - the distance the ray path travels

Basic geometry of a SAR satellite

SENSOR ORIENTATION

The SAR sensor is mounted on a satellite and points sideways instead of straight down like an optical sensor. This is because SAR sensors measure range (R). So if the ground was illuminated vertically down (nadir) there could be two points that measure the same range in different places at the same elevation. For this reason, SAR sensors are oriented in a side-looking geometry to ensure that those points for a given line would be placed in a different pixel location on the image.

The geometry of a side-looking SAR sensor (left) and a nadir looking optical sensor (right) showing how the range in two different places at the same elevation (R 1  and R 2 ) would measure.

DISTORTIONS

Radiometric and geometric distortions occur due to the side-looking imaging geometry of the SAR satellite. The figure below shows different types of distortions that can occur when the signal interacts with the terrain on the ground and are addressed in the subsequent sections. The recorded SAR sensor image is shown on the slanted plane and marked SAR image. The different colors indicate how the ground surface maps to the pixels in the SAR image.

Side-looking imaging geometry (Adapted from Pinel et al., 2014)

FORESHORTENING & LENGTHENING

The blue slope facing the sensor is foreshortened and the purple slope facing away from the sensor is lengthened in the SAR image. This means that the resolution is decreasing in the foreshortened region and increasing in the lengthened region. So, although both slopes are the same length, the slope facing away from the sensor has even more pixels than the slope that is facing the sensor.

LAYOVER

Layover occurs when the radar signal reaches the top of a tall feature before it reaches the base. The green section of the steep mountain is an example of layover, it appears in the same pixel as the ground surface.

LAYOVER & SHADOW

The brown section is a combination of layover and shadow, which appears to the right of the upper section in the SAR image, despite being located on the left of it on the ground.

SHADOW

Shadow occurs when an object blocks the radar signal. In the illustration, the yellow slope facing away from the sensor is not illuminated. Since radar illumination is not scattered in the atmosphere the shadows in a SAR image appear black. In the illustration, the yellow slope facing away from the sensor is not illuminated.

SPATIAL RESOLUTION

Active sensing enables a SAR sensor to synthetically increase its spatial resolution. A SAR sensor emits electromagnetic waves with a chirp of varying frequency that serves as a marker in the received waves. As a satellite orbits or an aircraft flies along its track, the SAR sensor images a point on the ground surface multiple times. The chirp marker is used to identify the location of the received waves. This feature, combined with signal processing techniques, enables a SAR sensor with a short antenna to synthetically elongate its antenna, which enhances its spatial resolution. To identify the location of the received wave on the ground, the SAR sensor must be side looking. If a SAR sensor is nadir looking (pointed straight down), it cannot use the travel time to distinguish between features that are an equal distance from the sensor on opposite sides.

Depiction of synthetically increasing aperture for a given sensor and using concepts from camera optics to analogize the processes.

CONTACTS

Esri Radio Detection and Ranging (RADAR) Specialists

Elizabeth Ashley Menezes, MSc

emenezes@esri.com Hello, World! Thanks for reading our StoryMap till the end! I am an Esri Raster Product Engineer on the Imagery & Remote Sensing team. I focus and work on UX/UI design and development of RADAR & SAR based capabilities at Esri. My education is in geophysics and my thesis work centered around understanding induced earthquake occurrences using seismic data and satellite imagery. As an advocate for science communication, I enjoy bridging the gap between science and design — combining my technical knowledge with my keen eye for design to create beautiful science

Magali Barba-Sevilla, PhD

mbarba@esri.com I am the Lead Esri Raster Product Engineer on the Imagery and Remote Sensing team working on RADAR for ArcGIS Pro. I just completed my PhD in geophysics at the University of Colorado Boulder. My research focused on synthetic aperture radar (SAR) applications in earthquake science. In addition to seismology from space, I am passionate about STEM diversity and outreach. I am the cofounder and former chair of the Society of Latinxs/Hispanics in Earth and Space Science (SOLESS).

Elizabeth Ashley Menezes

emenezes@esri.com

Magali Barba-Sevilla

mbarba@esri.com

An optical sensor (left) only works with clear, sunny skies. A SAR sensor (right) can work day and night, and even when clouds are present.

Electromagnetic spectrum showing short to long wavelengths and the respective types of waves (from left to right – gamma rays, x-rays, ultraviolet, visible, infrared, microwave, radio) as well as the different bands that SAR sensors can use are available within the microwave range.

Range of wavelengths for microwave bands categorized into short, medium and long wavelengths

Animation showing the sensor transmitting vertically polarized waves and receiving vertically polarized (VV) or horizontally polarized (VH) waves, as well as transmitting horizontally polarized waves and receiving horizontally polarized waves (HH) or vertically polarized waves (HV).

Basic geometry of a SAR satellite

The geometry of a side-looking SAR sensor (left) and a nadir looking optical sensor (right) showing how the range in two different places at the same elevation (R 1  and R 2 ) would measure.

Side-looking imaging geometry (Adapted from Pinel et al., 2014)

Depiction of synthetically increasing aperture for a given sensor and using concepts from camera optics to analogize the processes.