The new era for satellite SAR missions
The scenario faced by the SAR industry is unprecedented: the number of satellite SAR sensors is increasing steadily, as well as the number of radar data users, while the cost of both radar sensors and satellite imagery is getting lower. What is going on?
We spoke to Alessandro Ferretti, CEO and co-founder of the TRE ALTAMIRA group for the answer. Dr Ferretti has over 20 years’ experience in SAR data processing and SAR interferometry, being the inventor of the PSInSAR technique.
A golden age for SAR sensors?
To a certain extent, the SAR business is following the path of optical sensors: from extremely expensive missions, providing data for intelligence and security, and paid for by governmental agencies; to flocks of small, cheap, micro-satellites, providing data to an ever-increasing number of civilian applications.
How can we explain this new trend?
Actually, it is a paradigm that we often find when studying the history of many of today’s commonly used technologies (think of GPS, for instance). The proof of concept and the first prototypes are often fully paid for by governmental agencies. Business aspects and commercialization comes later, when ad hoc components and manufacturing processes are replaced by off-the-shelf products and standardized solutions.
How is the space segment changing?
Today it is possible to manufacture and launch a satellite SAR sensor at only a fraction of the cost that was associated with this kind of mission 10-15 years ago. Business plans involving SAR satellites are sounding more and more reasonable to investors, despite the high-risk factor still associated with the aerospace sector. Companies like ICEYE, Capella Space and Umbra Lab are evidence of this new trend. After all, recent analyses report that the earth observation (EO) business is expected to grow with the Compound Annual Growth Rate of about 14% over the next five years. There is a growing demand for EO data and people want to get that information independently of sun illumination and cloud coverage. It is a clear call for radar data that will complement more and more optical images in any monitoring solution based on satellite imagery.
So… SAR data is just a backup solution when optical sensors cannot provide any useful information?
Well, it is much more than that. Radar sensors see something different, compared to optical devices. First, because they are active sensors and work in the microwave domain, they are gathering information about a different set of geophysical parameters characterizing the area of interest. Second, more and more SAR sensors have polarimetric capabilities, which are extremely important for running automatic classification algorithms over wide areas. Finally, yet very importantly, radar sensors, being coherent, can acquire both amplitude and phase information, making it possible to apply interferometric techniques and get InSAR data.
What are the main applications of InSAR data? Do you see any trends?
InSAR data have two main applications: DEM reconstruction (think of SRTM and WorldDEM missions) and surface deformation monitoring. Both have a market, though InSAR applications related to displacement measurements are still a small niche in the EO business. Interestingly, despite the fact that the first results date back to the years immediately after the SEASAT mission (1978), this kind of data is still considered a “new technology” in oil & gas, mining and civil engineering projects. Definitely a slow uptake for a technology having immense potential: being able to detect from space a displacement of a few mm sounds like magic, but it can already be done today, using InSAR!
Why?
Apart from the lack of awareness about the kind of information that can already be obtained with the currently available SAR sensors, one of the reasons for this slow market development is related to the space segment. The most interesting InSAR applications require high temporal frequencies of acquisition over the same area of interest, using the same acquisition geometry and acquisition mode, i.e. short repeat cycles. To get decent spatial resolution and global coverage, the first SAR missions could acquire data stacks of just a few images per year over the same area. This was enough to monitor the creeping of a seismic fault or the slow inflation of a volcano, but not enough to provide useful information during, for instance, the excavation works of a new underground line. That is why future satellite constellations can be a game changer.
What kind of scenario will emerge?
In my opinion, in the next few years, space agencies and governments will continue to finance powerful (and expensive) sensors capable of acquiring data all over the planet with a resolution of a few meters and a temporal frequency of a few days. Small and relatively inexpensive satellites, most likely operating at X-band, characterized by high spatial resolution and short repeat cycles (<3 days), will be operated primarily by private companies. Their data will be used – apart from intelligence and security – for applications related to monitoring individual assets (e.g. mining areas, oil & gas reservoirs, etc.) and infrastructures (e.g. pipelines, lifelines, etc.). The more expensive machines will be used for wide area mapping, for monitoring natural hazards, climate change-related phenomena and urban sprawl, as well as for science applications.
What are the main challenges?
The main challenge faced by the operators will be the translation of this data flood into valuable information, easily utilized by end-users. This very point has already become a mantra within the EO community. People are realizing there is a need for both fast data processing through cloud computing, and the introduction of new algorithms and techniques for data screening, clustering, pattern recognition and automatic target classification. We are facing a paradigm shift: satellite imagery will become more and more affordable and the most important step in the value chain will be the last one: providing timely and reliable information to decision makers.