Issue #96 of Position magazine is hot off the press, and features the below interview with Airbus’ synthetic aperture radar guru, Ralf Duering.
A space-borne sensing technology that’s been flying under the radar for some time is now coming of age, and its maturation process has spawned a formidable set of products and services. Synthetic Aperture Radar (SAR) imaging features a host of different characteristics to optical imaging from space – it is impervious to weather and lighting conditions, but requires considerably more processing muscle and smarts to extract meaningful insight outputs. Position sat down with Ralf Duering, Asia-Pacific Business Development Director for Airbus’ TerraSAR-X platform, to discuss some of the ins and outs of this intriguing technology.
Position: So Ralf, why radar? Could you summarise the technology and its development to date?
RD: Radar works like a police speed check – a signal is sent out at a target, let’s say a motorcyclist, and as the signal returns a number of times, the target’s speed can be measured. SAR works the same way. TerraSAR-X was the first commercial radar service in space, and when we started out, the industry was not immediately aware of this technology and what it could offer. They were used to working with optical images. Over the years though, recognition grew – an understanding that this was the most precise measuring tool in space, which is how the NGA [The US National Geospatial Intelligence Agency], the American military procurement agency, classifies this sensor.
Position: Can you describe some of the ideal use cases for SAR, or recent applications? Under which circumstances does it offer advantages over other forms of imaging?
RD: One key technology is INSAR [Interferometric Synthetic Aperture Radar], which we use for highly precise change detection. Since the radar is completely in control of the signal – its phase is known – we can detect extremely fine variance on its return. Additional fractions of a wavelength added to the signal represent change in the targeted object. The created image is called an interferogram, made up of two images of the same target, and the measured phase difference between the returned waves representing change – surface change in the order of millimetres, for example, that we can monitor over time. Construction stability in civil engineering projects, sea ice movements, landslide monitoring and tracking land subsidence are key applications for this technique. Currently the Chinese government has 67,000 bridges across the country that they need to monitor for structural issues in this way. 67,000.
Time series analysis with SAR also has key surveillance applications, as you might imagine – maritime monitoring, of illegal fishing in Indonesia for example, construction of facilities, vehicular movements. We can do very, very nice automatic ship detection because of the dielectric properties of the vessels against the ocean, which we can do with coherence change – something you can’t do with amplitude changes from optical images. Even with 40-metre resolution, you can easily identify 12 to 18-metre ships, even wooden ships, because of the way the wave bounces on the water and comes back. We provide a lot of these services in this area to Southeast Asia, monitoring illegal deforestation in PNG, Malaysia, Indonesia, and construction sites in China.
We recently created the World DEM [Digital Elevation Model], a global 3D elevation model, including the north and south poles, with 2-metre relative accuracy.
Stay up to date by getting stories like this delivered to your home or office:
Subscribe to Position magazine
RD: Finally, what we can offer in terms of ground control points (GCPs) is unique. This service has historically only been applied in inaccessible or denied areas, but we’ve now reached a domain within which we can now extract GCPs automatically with 10 to 20 centimetre accuracy from space, and up to one metre x, y, z. This could change the game for calibrating street maps, and holds massive potential for autonomous driving mapping applications. This one is a new business case even for us, we only found out in the last few weeks that we could do this – step by step, this technology has really evolved. We had thought that this service would only be in demand for the autonomous vehicle space, but we’re seeing many clients scrambling to take stock of this development – Google, Beidou and others. We anticipate demand to develop beyond the autonomous vehicle street maps case quite quickly.
Position: My understanding is that the Spanish PAZ satellite that was launched in February is part of your constellation, but features an AIS [Automatic Identification System] module along with an SAR sensor. Was there a specific business case for including this unit?
RD: PAZ is an exact clone of our TerraSAR-X and TanDEM-X satellites, with the only difference being the AIS on board. Ship detection is one of our primary applications. Often when ships are engaging in dumping oil or other illegal activities, they’ll switch off their transponder, or send out a false signal – perhaps indicating they’re 150 kilometres from their actual position. So what we do is find these ships, check if they’re sending a signal, and whether it’s a proper or false one – and then we can go hunting. Indonesia is one of our biggest clients in the region we work with SKK MIGAS [The Special Taskforce for Upstream Oil and Gas Business Activities] to monitor all these straits in their waters. It’s very obvious and quite shocking to witness where they dump their oil when cleaning their tanks – it’s a huge stripe that goes out 200 nautical miles, and it just stays there. So with the AIS receiver on board, we don’t need to collect the signal, we can instantaneously compare it with the image and see what’s happening.
Position: Surely it’s not all roses though, what are some of the drawbacks?
RD: Well, unlike optical, radar satellites take six months to calibrate, verify correct system configuration and the validity of data outputs. So around six months after launch are the services operational commercially, so we’re expecting the additional revisits offered by PAZ, and its specific service offerings to be live in about December. This is versus about a month and a half for optical, which is basically already there and working once in orbit. Observation of large areas can be better achieved with optics in many cases, depending on your requirements.
We went through a huge learning curve on the processing, to be frank. We started using radar like optics, initially. We knew the technology was capable of producing such precise measurements, but it took time for us to really figure all of this stuff out. Processing power is no issue anymore, it’s developing and optimising the algorithms that takes the time – there’s a huge difference between understanding what’s technically possible and having a published paper that tests it, and physically making it reliably applicable for commercial use. But four to five years in, and we’re there – and still discovering new benefits.
Stay up to date by getting stories like this delivered to your mailbox.
Sign up to receive our free weekly Spatial Source newsletter.
Position: Are there any specific benefits of this technology that can assist with survey-related applications, for those out in the field?
RD: I think some surveyors have been afraid in the past that we would take their jobs. Because we can easily generate such high point density and offer such extreme accuracy in regards to surface movement – and everything constantly moves. But there’s actually quite a lot we can offer here, in terms of optimising levelling and GNSS measurements and instruments. Because very often, when they go on construction sites, on mining sites, they need to put a grid in, artificially – as to where they think they should monitor. But they also monitor positions, locations where there’s no need to – it’s stable ground. We saw it at a mining site with subsidence issues just this morning – total stations surveying equipment set up in many positions in which there was no need, whereas just a hundred metres or so left and right, whole walls were moving, and nothing was set up there. Of course they couldn’t know this, but with access to some of our products, we can help give this hint on where to look, and then they can monitor every second in the areas that need it.
Position: What is that service you refer to, how can surveyors access it?
RD: We typically see mining companies purchasing these products, usually under their corporate responsibility codes or as a preventative safety measure. So they buy the data, and it’s an internal decision as to whether they make it available to their surveying team, but we want them to use it, we encourage this, because there are a range of other environmental benefits that flow from this as well. We’d rather not see it just sitting on the office of the geotechnical people on site and the surveyor’s don’t get to use it.
Position: Thanks for your time today, Ralf.
RD: You’re welcome.