Satellite Data

Satellite data or satellite imagery is understood as information about Earth and other planets in the space, gathered by man-made satellites in their orbits.The most common use for satellite data is Earth Observation (EO): satellites deliver information about the surface and weather changes on the planet Earth.

Satellite data is generated via remote sensing technologies. One of these, Synthetic Aperture Radar (SAR) technology, provides reliable and timely data stream opportunities for a myriad of applications at land and sea. For governments and organisations, this means possibilities for acting on time for example in connection of natural catastrophes and crimes, for businesses this enables developing completely new and improved services.

One of the inherent value that satellite data has is the availability of authentic information about Earth surface, weather, and other incidents. Cumulative data helps to understand developments in long-term and timely data to act swiftly on the detected issues. The number of satellite data applications is unlimited, providing manyfold advantages on global and local scale.

1.1. EARTH OBSERVATION TECHNOLOGIES IN BRIEF

Earth Observation (EO) refers to the use of remote sensing technologies to monitor large areas of land, sea, and atmosphere. Remote sensing means observing an object (in this case the Earth) with the help of device which is physically remote from it.

Remote sensing devices with EO sensors can be carried with various platform groups:

• Spaceborne, meaning satellites, shuttles or space stations
• Airborne, meaning manned or unmanned aircraft, aerial vehicles etc.
• Ground-based, meaning installed sensors, mobile devices, animals or humans

Most of Earth Observation data comes from spaceborne platforms with remote sensing tools. Earth Observation through satellites relies on the use of their mounted payloads which collect and transmit images or other data of the Earth’s characteristics. The Earth Observation satellites are normally launched to relatively low altitude orbits (about 600-800 km from the surface of the Earth) to be able to carry out their imaging tasks. 

The fundamental differences of EO technologies depend on whether the data is collected through a passive or an active remote sensing system.

Passive remote sensing
A satellite has a remote sensing payload to monitor the solar energy which Earth’s surface or atmosphere reflects and re-emissions into space. This technology delivers data only in sunlit, cloudless circumstances. Commonly used passive sensors include optical and thermal sensors.

Active remote sensing
A satellite sends energy to Earth and then measures the energy received back from the Earth’s surface or atmosphere. Commonly used active sensors include radar and laser technologies. There are two types of radar-based systems that are used for satellite-based microwave imaging:

• Real Aperture Radar (RAR)
• Synthetic Aperture Radar (SAR)

Active imaging systems, especially Synthetic Aperture Radar, is the only instrument that can penetrate clouds and doesn’t require sunlight to provide reliable remote sensing data. This is especially relevant for monitoring those areas of the Earth which most of the time are covered with clouds, such as in Northern Europe, South American rainforests, or South-East Asia.

2.1. OPTICAL SATELLITE IMAGING

Satellites with optical sensors represent passive remote sensing technology. They only can generate images of those parts of the Earth that reflect sunlight. In dark or cloudy weather optical imaging is not possible, or returns imagery of bad quality. Optical data includes the visible wavebands and produces images that are similar to the way how the human eye captures the world. In addition to visible light, an optical sensor captures light in the near and middle infrared, that is not visible to the human eye but which contains important surface characteristics information. The resulting reflectance spectrum is analysed afterwards.

A passive sensor records the radiation of its target within a solid angle of view. Near, short and middle infrared wavelengths suit for distinguishing vegetation types with moderate resolution, where pixel size varies from 10 x 10 to 30 x 30 meter. Optical satellite data can be combined with laser data which provides additional information about terrain or vegetation characteristics.

In addition to the actual reflectance of the target surface, the radiance detected by an optical sensor depends on many factors including atmospheric conditions (clear or dusty atmosphere), direction of illumination source (Sun angle), scattering of the target (rough/ smooth surface), orientation of the surface (influences topographic shadowing and/or increased radiation on rolling ground), and view angle of the sensor (nadir looking or off-nadir looking) (CRCSI 2017).

2.2. SYNTHETIC APERTURE RADAR (SAR) IMAGING

Synthetic Aperture Radar (SAR) satellites enable imagery to be obtained during both day and night-time. In addition, SAR can penetrate clouds and it works in all weather conditions, including rain and snowfall. SAR sensor carries its own illumination source, in the form of radio waves transmitted by an antenna.

"Synthetic aperture" means physically moving the small SAR instrument over an area, while it gathers information of its target, rather than having a very large instrument which would image the same area at once. To achieve this movement, SAR instruments are often mounted on satellites and airplanes, which enables them to cover large areas in a very short time.

SAR synthetically increases the antenna's size to increase the azimuth resolution through the same pulse compression technique as adopted for range direction. A shift in wave frequency is known as the Doppler effect. The distance between the Satellite radar antenna and different reflection points on the Earth's surface are measured and processed into images readable by humans and machines.

Active remote sensing systems operate in the visible and microwave parts of the electromagnetic spectrum. SAR measures microwave response from surface features, potentially by using multiple polarisations. This creates reliable imaging of both land and sea. Imaging can be used for vital decision-making at public and private sectors, where it helps to monitor sea ice, marine or land environments and events, and support acting on crises requiring humanitarian aid.

Due to the nature and variety of observed surfaces, human-made structures are more likely to be visible on SAR images already at a lower resolution. The amount of returning reflection from surfaces is called backscatter. A SAR image is essentially a square raster-based output of the measured backscatter as an image.

Reflectance from a surface is generally not uniform since it varies with the texture of the surface relative to the wavelengths of incident radiation. Reflectance can vary from specular (totally smooth) to diffuse (totally matt), but most surfaces display characteristics between these two extremes. A smooth surface acts as a mirror and is seen as black in the SAR image since the echoes never reach the antenna. For example, ships are seen very well on the surface of the sea.

Read the story of this ICEYE-X2 SAR image by Michael Sheetz from CNBC

2.3. RADAR BANDS AND POLARISATION

Radar band, in essence, is a selection of wavelengths or range of radar frequencies. Most commonly used SAR bands are X-band, C-band, and L-band. Each radar band has its own penetration characteristics which are affected by different materials and conditions on the ground.

X-band

With a nominal frequency range from 12.5 to 8 GHz (2.4 to 3.75 cm wavelength) within the microwave portion of the electromagnetic spectrum, X-band is a suitable frequency for several high-resolution radar applications. It has been applied for both experimental and operational airborne and spaceborne SAR systems, designed for various remote sensing applications. The corresponding wavelength for these systems is around 3 cm, which has proved to be useful for mapping and monitoring assignments.

C-band

This band has a nominal frequency range from 8 to 4 GHz (3.75 to 7.5 cm wavelength) within the microwave portion of the electromagnetic spectrum. C-band has been commonly used for several experimental aircraft SAR systems in addition to several single-band satellite SAR systems. The corresponding wavelength for these systems is on the order of 5.6 cm, which proved to be useful in sea ice surveillance and other applications. Like other radar bands, C-band allows to see even through the thick layers of clouds. Its penetration characteristics of vegetation canopies or soils are limited and are only possible for the top layers.

L-band

The nominal frequency range of L-band is from 1 to 2 GHz, (30 to 15 cm wavelength range). The L-band is one of the main operating ranges used by various applications in addition to radars, such as Global Positioning System (GPS), radio, and telecommunications. The L-band has a low bandwidth because of its low frequency which allows also for easy implementation. For example, some materials such as snow and sand can be penetrated to a degree by microwaves. L-band penetration is very consistent, supporting applications such as underground archaeology.

Polarisation

The scattered signal amplitude and polarization provide information about scattering properties and structure. Phase measurements of the return pulse are used in interferometry and other applications. In vegetation canopies penetration example, X-band radar, with wavelengths of around 3 cm, is scattered strongly by foliage, while L-band, with wavelengths of around 27 cm, is scattered more strongly by tree branches.

Mirror reflection occurs for smooth surfaces with a low amount of radiation returning to the sensor. In contrast, rough surface will cause for a bigger amount of radiation to backscatter. If solid targets, such as metal objects, are oriented towards the antenna, they will have a very strong return signal. Amplitude is recorded always, while advanced systems can also measure polarisation and return signal phase.

Imaging radars operate in one or more of the select polarisation modes. Radar system measurements can be single-polarised (HH or VV or HV or VH), dual-polarised (HH and HV, VV and VH, or HH and VV), or quadrature polarised (HH, VV, HV, and VH).

Co-polarised measurements

• HH — horizontally polarised signal, horizontal component of the return signal is measured
• VV — vertically polarised signal, vertical component of the return signal is measured

Cross-polarised measurements

• VH — vertically polarised signal, horizontal component of the return signal is measured
• HV — horizontally polarised signal, vertical component of the return signal is measured

For polarimetric analysis, signal phase differences between the modes and signal magnitude are measured by polarimetric (quadrature polarised) radar and some dual-polarised systems. Orientation and shape of the object surface affect the polarisation of reflected microwave energy in varying degrees. The final appearance of the objects in the images is affected by the polarisation mode trough scattering intensity echo.

Synthetic Aperture Radar satellites collect imagery during day and night, in all weather conditions.SAR satellites receive return pulse and based on different angles of this pulse, it is possible to create 3D images. Revisiting the same places in their orbits, SAR satellites create the same image quality regardless of snow and rain. This provides a significant advantage especially for change detection applications at land and sea throughout the globe.

ICEYE has developed innovative radar technology for under 100 kg microsatellites. This technology and optimal resolution choices enable dramatically improved cost-effectiveness over traditional large-size radar satellites.

3.1. SPECIFIC ADVANTAGES OF SAR IMAGING

SAR imagery provides users with increased reliability thanks to its all-weather functionality. In addition, radio bands allow a certain degree of object penetration which helps to identify meaningful features and properties of the Earth and its objects. True characteristics can be determined from SAR imagery, which can only be visible in the radar imagery, such as vegetation, forest canopy, surface roughness, and ice thickness. In general, radar remote sensing advantages include penetration through the vegetation canopy and soil, as well as sensitivity to structure and dielectric properties (liquid vs. frozen water).

The key properties of SAR imaging are:

• Independence of light and weather conditions, which enables delivery of satellite data any time irrespective of the conditions in the Earth atmosphere
• Ability to monitor large areas
• Ability to show extremely well surface conditions of large sea, lake and river areas as well as detect objects at the vast sea areas
• Ability to demonstrate movements of the objects especially on sea

While SAR satellite imagery seems to provide interesting opportunities for a specific use case, the first step is to consider in more detail if the technology is applicable and suitable for the task at hand. A reliable partner helps to evaluate how SAR imaging stream or imaging tasks help to create actionable information and what kind of data combination, verification, and analysis services are needed.

SAR imaging technology enables better grounds for decision-making thanks to reliability and frequency of the satellite data, sufficient coverage and resolution of the images as well as easy-to-apply data integrations with the existing operating systems.

4.1. REFINING SATELLITE DATA INTO ACTIONABLE INFORMATION

The process of turning satellite data into useful information takes place in several steps. As satellites can be equipped with different type of sensors, the data content of the acquired images is different, too. Consequently, the raw data source, analyzing tools, and processes needed for turning the imagery into valuable information must be chosen carefully, starting from the purpose and end use of the information.

Recent advances in analysis methods have opened a wide door to applying machine learning and artificial intelligence in turning satellite imagery into actionable information.

1. Planning of data acquisition (area, image features, frequency, timing)
2. Collecting images and downlinking them to ground station
3. Processing of imagery for analysis
4. Analyzing data (when needed, combined with data from other sources)
5. Information delivery (for example. information visualizations, full reports)

The acquired satellite imagery is first downlinked to a ground station and prepared for analysis. This stage can include steps such as lining up pixels, removing clouds and mist, polar reformatting (for SAR data), and so on, depending on the type and end use of the satellite data.

In many use cases, satellite data is analyzed together with data from other sources. For example, a Geographic Information System (GIS) analysis can combine optical and radar data, digital elevation models, and digital orthophotos, providing an analytical foundation for modeling and feature extraction. Satellite imagery has raster (cell-based) layers which can be used to map and model virtually anything that happens on the Earth. Such cell-based data structure enables systematic and controlled analysis with multiple data layers.

The goal of any style of information design is to highlight what is significant in the acquired data. Visualization of satellite data improves our understanding of the significance as it is often difficult to understand what massive amounts of numerical data really tells us.

Here are some examples of satellite data analyzing methods and their applications:

Change detection analysis

Information on how a selected area has changed between two or more time periods is helpful for understanding and evaluating, for example, forest coverage, ice sheets, and urban land use. Satellite imagery can be used to automate the classification and locations of land into specific categories, such as different types of land uses and land cover. Image differencing and change dynamics analysis help to evaluate past changes in relation to the present, as part of assessing the accuracy of the change detection technique.

Image segmentation analysis

Segmentation of satellite images enables various monitoring tasks, from movements of populations in remote areas to evaluate the potential roofs for solar panel installation. Each pixel in the data is classified and grouped with similar pixels and an average value is assigned to the grouped pixels. Machine learning models based on segment analyzed satellite images and ground-based data can automatically detect objects of interest on the Earth’s surface.

Suitability analysis

Suitability analysis using GIS and satellite imagery improves the correctness of land use classification and urban structure analysis. The information is used for selecting suitable places for building or preservation, for agriculture, and so on. Site selection analysis can be performed with vector or raster data.

Visibility analysis

Visibility analysis is used mainly for land use planning purposes. Visible information of a location can be used, for example, for determining the value of a real estate and for selecting locations of telecommunications towers. An output raster displays what entities can be seen over a certain landscape, from the point of view of one or more objects. Simulation tools help to create a 3D environment using DEM files and shapefiles of the targeted objects (like buildings).

Increasing availability of satellite imagery influences many global and local processes and industries.Accurate information helps to understand how the world is changing on a day-to-day and hour-to-hour timescale. SAR imagery proves changes in long-term and shows what is going on right now.

The applications vary from large scale to very local issues: for example from natural disasters to identifying sea traffic in a selected area. The applications can be categorised for example, to land and maritime end uses.

For maritime use cases, SAR satellite data provides three main benefits compared with optical data:

• Ability to gather radar imaging data through clouds and darkness., which means reliability
• High contrast between water and other elements on the surface of the water (such as vessels and oil), meaning an increased chance of correct, useful findings
• Ice monitoring capability due to radar wave penetration

For land use cases, SAR satellite data provides these main benefits compared with optical data:

• Imaging in all weather and light conditions, meaning reliability
• Surface roughness determination
• Vegetation canopy penetration to a certain degree

5.1. LAND APPLICATIONS OF SAR IMAGING

The wellbeing of human race largely depends on the state of the Earth resources. In order to be effective in managing land resources and rapidly reacting to changes that occur around us, governments and corporations need to have access to a stream of reliable data, which will not be subject to atmospheric changes. According to NASA, about half of the total land area is usually covered with clouds. For example, there is a particularly cloudy zone over the equator which is very often covered by thick clouds, preventing optical remote sensors to picture the land below them regularly.

The land applications of SAR satellite data range from agriculture monitoring and forest mapping to urban infrastructure surveys. The information is applicable for governmental, non-governmental and business specific needs. The most important applications are

• Security surveillance
• Agriculture harvest monitoring and field segmentation
• Mapping of areas affected by natural disasters and NatCat response
• Infrastructure and construction monitoring
• Oil pipeline monitoring, oil tank fill rate
• Interferometry (ground deformation)

Changes in field harvest detected from ICEYE-X1 imagery (featured in Quartz)