Subterranean Mapping

_{(5 min read)}

Let’s start with New Year’s resolutions and try to tidy up the data about our environment. We often talked about the need of different stakeholders to put together their data and make it usable; we also talked about representing the roads and rails we are traveling on. But there is more essential infrastructure for our modern, daily life: supply and disposal infrastructure. In our Western world we often associate a modern state of the art with the existence of proper communication, transport and supply infrastructure. If a location has no power and water connection at all it is underdeveloped. Thus, in the last centuries we built a lot of underground infrastructure, but it is way more complex to get the full picture about our environment below the surface. A first good overview can already be gained by looking at OpenStreetMap because it contains a lot of infrastructure but without displaying it in a prominent way. The Open Infrastructure Map is exposing this data with focus on powerlines and powerplants, solar plants, oil, gas and water pipelines and corresponding facilities, and communication and shows also the footprint of artificial illumination. But if you want to know the details Open Infrastructure Map scratches only on the surface of the underground.

Since 2022 the Geospatial Commission, part of the Department for Science, Innovation and Technology of the United Kingdom, is building a digital map of underground pipes and cables in England, Wales and Northern Ireland – the National Underground Asset Register (NUAR). We take this opportunity to dig a little bit into the underground.

Currently the minimum viable product (MVP) of the underground register is still under development and the NUAR platform shall be fully operational by the end of 2025 (including a public beta for eligible users in spring 2025). Currently the “safe dig” use case is supported where you access the platform and generate a “work pack” for your construction site with the data of involved stakeholders (one PDF export per asset owner). It shows the known infrastructure elements and their features such as asset owners, protection state and additional temporal information (planned activities such as maintenance).

Various stakeholders are involved in underground infrastructure. Multiple companies could be in charge of sewage, electricity, communication, gas supply, district heating, supply shafts, transport tunnels, substructures and ancestors infrastructure (or old military equipment) at the same location. Aforementioned NUAR already has to deal with more than 600 public and private sector asset owners, which operate 4 million kilometers of cable and pipes.

In previous times everybody kept their documentation for themselves, and often documentation about underground structures got lost. When a new construction site is set up it sometimes causes accidental strikes on underground pipes and cables, which leads to power supply disruptions. Bringing together the data seems to be the solution to know everything about the site where digging is planned. But accidents can’t be avoided completely because often some of the infrastructure is poorly documented or even not documented at all because it was built decades ago. Currently around 60,000 times a year accidents such as cutting power lines or – even deadlier – cutting gas lines are happening, causing costs of up to 2,86 billion EUR.

By the way: Knowing where infrastructure is located can also be misused. In recent times infrastructure such as railway signaling as well as power supply and data connections got more and more attacked. But hiding the information is only security by obscurity. The above-mentioned NUAR is providing access only by invitation, which is initiated by asset owners. Additionally all accesses are monitored and documented but data leaks or security vulnerabilities could change the situation.

Nevertheless, data fusion is necessary to get the full picture independent of whether a large-scale (e.g., multiple deep basements or tunnels) or a small-scale project (e.g., installation of new network cables) is to be realized. As of today, you have to ask every data owner to get related data and after that fuse the data, which very likely will have different levels of detail and scale in three dimensions. Only then may it be the case that you get a good overview of documented infrastructure.

But how to represent the data? Representing an overground environment is kind of easy: We already have various data formats to represent road, rail and the rest of the environment (or everything together). The underground infrastructure can vary from cables via sewers to large-scale basements. For that the NUAR collaborates with the Open Geospatial Consortium to develop and publish an international standard for describing sub-surface assets called Model for Underground Data Definition and Integration (MUDDI). NUAR is the first implementation of this data model using OGC Simple Features (point, line and polygon for areas as well as everything in “multi” for volumes resp. bodies) to represent the objects so that it gets easy to merge and compare the data with other cadastral information.

Data acquisition above ground can easily be done with the help of mobile mapping and analysis of aerial images. Various companies offer services and provide different levels of detail. But looking below the surface is more complicated and requires the third dimension. A good example is the layout of public transport underground stations, very nicely visualized by Albert Guillaumes Marcer.

In addition to all of these infrastructure items, the “medium” in which they are implemented, matters, too. The ground conditions (e.g., sandy soil, brown coal deposit and aquiferous stratum) can be relevant for design decisions especially for large-scale constructions. Measuring the soil layers is not as easy as “looking around”, especially if the surface is already overbuilt by other infrastructure.

All this infrastructure and foremost the different layers of soil generate an individual fingerprint of every location in the world. Years ago there was the idea to use this fingerprint for localization esp. on roads because this method would be independent from bad weather situation (e.g., snow cover) or can be utilized also for indoor applications. With the help of a ground-penetrating radar the underground of the roads get scanned and the sensor response is transformed into some kind of hash value that can be easily stored in a lightweight database. If you scan the road a second time this hash value can be tracked down in a corresponding dataset and you know your location without the need of using a GNSS or other landmarks. Due to the fact that reality changes only slowly, the dataset stays up-to-date for a long time. The drawback is that the ground penetrating radar sensor kit is kind of expensive and, additionally, there is extra effort for creating the dataset compared to GNSS, which is ubiquitously available (the positions of the satellites can be calculated with the help of an almanac).

Ten years ago TomTom launched a similar solution utilizing the sensor sets in current vehicles instead of introducing a new one and called it RoadDNA. The idea was to scan the right and left roadside and represent the 3d depth map as simplified 2d raster data. But this idea was discontinued because of the effort to create and maintain the dataset (remember, the environment above the ground changes more frequently than the underground). At the end also the ground penetrating radar localization has only limited purpose if at all. Nevertheless, building a digital twin of our increasing underground infrastructure that is additionally also getting more complex is necessary to manage our modern life.