Building a better future with a digital twin of the world

Many companies claim to be building a digital twin, but in reality, no one has yet built a true digital twin of the physical world. Ecopia is on a mission to build this digital twin, providing organizations across industries with a virtual representation of the real world for mission-critical mapping, analytics, and decision-making. So far, we’ve built a foundational layer for a digital twin of the world using up-to-date geospatial information and our AI-based technology.

The first step in creating this foundation of a digital twin is sourcing high-resolution geospatial imagery from our global partner network. With this detailed imagery, we then apply our AI-based mapping systems to extract all land cover features (both natural and manmade) with the accuracy required for a true digital twin. From there, we classify land cover and append height attribution where applicable to provide real-world context. Finally, our global partnership network supplies us with the most up-to-date imagery so our AI can detect change over time and digitize new or altered features.

There are many applications for the complete, accurate, and up-to-date data that forms the foundation of a digital twin, spanning from the public to private to nonprofit sectors. But while the future of mapping lies in this creation of a digital twin, it cannot be realized without the availability of high-resolution geospatial imagery. By joining Ecopia’s partner network, imagery providers, mapping product builders, and data resellers can empower their clients with detailed geospatial information while increasing their own revenue.

The future of mapping

Geospatial data and mapping are becoming increasingly important for mission-critical workflows across a diverse set of industries and use cases. While mapping used to largely be its own industry, the increased availability of geospatial data and accessibility of geographic information systems (GIS) tools has led many industries to bring their mapping and spatial analytics in-house. It’s never been easier for organizations to derive important information from maps.

However, despite this increased accessibility, complex challenges still exist when working with geospatial data. For this data to be useful, it must be comprehensive, accurate, and up-to-date. If the information captured in the data is incomplete, incorrect, or stale, any maps and insights derived from it are obsolete, and can actually lead to costly errors in decision-making. 

To avoid these pitfalls, mapping and geospatial experts are looking to the future and the idea of building a digital twin of the world. A digital twin is a virtual representation of the real world, always remaining up-to-date and capturing details with accuracy. Organizations can use a digital twin as a source of truth for structures, conditions, and activities in the physical world, ensuring that any maps or analytics generated from it are providing correct information that can be trusted for important decision-making. 

In this blog, we explore the possibilities unlocked by a digital twin, and how high-resolution geospatial imagery and high-precision vector maps are critical elements to developing a virtual source of truth for the real world. 

How to Build the Foundation of a Digital Twin

What companies need to build a true digital twin of the world

Because it represents reality, the first input needed to build a digital twin is an image of the real world. There are many different ways to capture geospatial imagery, such as from satellites, planes, drones, or even street view cars, and all can be used as the base layer of a digital twin. The most important thing is that the imagery is clear, shows the entire area needed in full detail, and is fresh. 

The next step is extracting features from the imagery so that they can be mapped and analyzed. While geospatial imagery shows different features on the Earth’s surface, GIS analysis requires the ability for these features to be isolated and appended with information. For example, while seeing a river on a map is helpful, what makes it really useful for decision-making is digitizing its exact extent and adding any attributes that may be needed for analysis, such as its name or width. 

Once each real-world feature has been extracted into a 2D planimetric-level map, height or z-axis attribution is added to make the data 3D. This is a critical step because to effectively represent reality, all features must be virtual duplicates of the real world, and real-world features are not two dimensional. With height attribution, data can form a true foundation of a digital twin and enable analysis that was previously only possible on-site or in-person. 

But building a 3D map is not the final step in creating a digital twin. In truth, a digital twin is never “finished,” because the world is always changing, and by definition a digital twin will always reflect that change. To detect change and maintain a digital twin of the real world, imagery must constantly be refreshed, analyzed, and digitized. However, building a detailed 3D map does lay the foundation for this digital twin, ensuring any insight derived from it is based on reality.

The Importance of Data Quality

The risk of using inaccurate data

Data professionals are familiar with the phrase “garbage in, garbage out,” which refers to how performing analytics on bad data leads to inaccurate results that cannot be trusted for decision-making. Given the process we outlined above, there are many ways stale, incomplete, or incorrect data can impact a digital twin, especially if it lies within the foundational layer.

If the geospatial imagery used as the base of a digital twin is of low quality, it is difficult to interpret features and extract them into individual map elements with the level of detail required to represent reality. For example, low-resolution input imagery can result in buildings being obscured or indistinguishable from their surroundings and ultimately being omitted from the output map. The final product is not a reliable foundational layer because it is incomplete and does not represent all features present in the real world. 

Similarly, using poor quality data as the foundation of a digital twin can cause features to be digitized incorrectly. Even if all real-world features are extracted, if the input data is too low quality to determine exact boundaries or characteristics of each feature, it’s impossible to build a digital twin. The resulting maps might be “close enough,” but they will not fit the definition of a digital twin - a virtual representation of reality.

Even the final step in curating a solid foundation can be derailed by low resolution geospatial imagery. To keep a digital twin up-to-date with a constantly evolving world, changes in both the natural and manmade environment need to be detected and digitized. If the imagery is stale, the previously mentioned issues of accuracy and completeness are compounded by input data being out-of-date. 

Any analytics performed on a digital twin with an inaccurate, stale, or incomplete foundation cannot be relied upon for decision-making. While it may seem that all of the information is there, key components may be missing that will lead to skewed results and costly mistakes.

The only way to be sure that a digital twin is really representative of the real world is to use trustworthy, fresh, high-resolution geospatial imagery as the foundation.

Applications for a Digital Twin

How high-resolution imagery is being used to build digital twins across industries

While no one has yet succeeded in creating a complete digital twin of the world, high-resolution imagery and AI are powering the way there. Many organizations are already leveraging high-resolution imagery and extracted vector maps for a variety of solutions to some of the world’s most complex challenges. In this section, we summarize a few of these applications to highlight how the highest quality geospatial imagery is bringing multiple industries closer to the goal of a true digital twin.

Example 4: Airbus’s 30-50cm resolution satellite imagery was used by Ecopia to digitize 3D buildings in Busan, South Korea for Snap Inc.

How Ecopia is partnering with imagery providers to build a digital twin of the world

Digitizing features from geospatial imagery is a standard GIS practice, but one that traditionally required many hours of tedious manual work. To extract a feature, a user would upload an image file to their GIS, then trace the outline of the feature and add any necessary attribution, such as what type of land cover it represented. If this sounds tedious for just one feature, imagine how much effort would be required over a large geographic area.

Ecopia’s AI-based mapping systems eliminate this need for manual digitization, extracting thousands of square kilometers of real-world features a day with the level of precision you’d expect from a trained GIS professional. To give you an idea of just how efficiently we can map at scale while maintaining parity with the real world, we mapped all buildings and roads of Sub-Saharan Africa in eight months; previous manual digitization of these features in just one country took GIS experts nine months. Similarly, we worked with our partner Woolpert (formerly AAM) in Australia to map all of Perth in 3D in just 12 days - a project Woolpert estimates would have taken 50 weeks to complete manually.

Using high-resolution geospatial imagery from our global partner network, Ecopia is digitizing the world and building the foundation for a digital twin of the Earth. Our maps empower not only our own customers, but also those of our partners. Many imagery providers work with Ecopia to offer their own clients more value with fully digitized maps, ultimately generating more revenue per pixel. Organizations building their own mapping products and applications also partner with us to digitize their source imagery at scale, resulting in high-precision vector maps their end users can trust for decision-making.

Another advantage Ecopia provides to our partners in addition to the highly efficient and accurate digitization of imagery is the affordability of our extraction services. Because high-resolution geospatial imagery is extremely detailed and the associated data files are so large, most other data extraction services charge more for digitizing them than low-resolution images. Ecopia charges the same flat fee regardless of imagery resolution, empowering our partners to help develop the foundation of a global digital twin with their best imagery.

To learn more about how Ecopia’s global partner network is building the foundation of a digital twin, reach out to our partnerships team.