The ability to fully harness 3D data is rooted in acquisition and scalability
The application of location intelligence and the incorporation of 2D maps and positioning have become ubiquitous since the advent of smartphones. Now, we are entering a new era in which we can harness the power of 3D data to improve consumer experiences, as well as applications for enterprise, public safety, homeland security, and urban planning. 3D will play a more significant role in these experiences as we overcome the technical barriers that have made it previously difficult and cost-prohibitive to acquire, visualize, simulate, and apply to real-world applications.
Outdoor Data: Our World Is 3D
In a geo-enabled community in which we strive for more precise answers to complex spatial intelligence questions, traditional 2D correlation is a limiting factor. When you think about 3D data and maps, modeling buildings in an urban environment seems obvious. However, 3D is incredibly important when trying to understand the exact height of sea level or the uniformity of roads and runways. For example, one can imagine the vast differences in 2D versus 3D data and its application during the 2017 hurricane season. By including the Z-dimension in analysis, we can achieve true, precise geospatial context for all datasets and enable the next generation of data analytics and applications.
Access to 3D data for most geospatial analysts has been limited. Legacy 3D data from Light Detection and Ranging (LiDAR) and synthetic aperture radar sensors has traditionally required specialized exploitation software, and point-by-point stereo-extraction techniques for generating 3D data are time-consuming, often making legacy 3D data cost-prohibitive. Both products cost hundreds to thousands of dollars per square kilometer and involve weeks of production time. Fortunately, new solutions provide a scalable and affordable 3D environment that can be accessed online as a web service or offline for disconnected users. Users can stream, visualize, and exploit 3D information from any desktop and many mobile devices. Models of Earth’s terrain—digital elevation models (DEMs)—are increasingly used to improve the accuracy of satellite imagery. Although viewed on a 2D monitor, DEMs deliver the magic through a true 3D likeness for bare-earth terrain, objects like buildings and trees, contours, or floor models, and unlimited contextual information can be applied to each measurement. This provides a true 3D capability, replacing current “2.5D” applications that aim to create 3D models out of 2D information, at a cost point closer to $10 to $20 per square kilometer and only hours of production time.
Indoor Data: Scalable 3D Acquisition
As 3D becomes more common in outdoor applications, its use for indoor location is being explored. Unsurprisingly, similar challenges need to be overcome. Until now, real-time indoor location intelligence has been difficult to achieve. This is largely due to the absence of, or difficulty in obtaining, real-time maps and positioning data to form the foundation for the insights businesses can derive about their spaces.
To create a 3D model of a room, businesses stitch together blueprints, photos, laser scans, measurements, and hand-drawn diagrams. Once the maps and models are in hand, operators must create location and positioning infrastructures that accurately track people and things within the space. Operators need to position these sensors—typically beacons—precisely according to the building map, but the beacons have no knowledge of the map and cannot self-regulate to reflect any changes to the environment. If the beacon is moved, its accuracy is degraded and its correlation to the map breaks down. Overall, this process is lengthy, cost-prohibitive, and fraught with error.
Using current methodology, professional services teams work with a wide variety of tools—LiDAR trolleys, beacons, 3D cameras, and existing architectural drawings—to compose an accurate representation of a space. Additionally, the resulting system becomes difficult to maintain. Physical spaces are dynamic, and changes quickly render maps obsolete. Changes to technology used to create the models or track assets require ongoing management, and the process rapidly becomes overly complex. This complexity is stalling innovation across myriad industries, including training and simulation, public safety and security, and many consumer applications. While organizations and consumers can easily find data for outside locations using GPS, no equivalent data source exists for indoor location data.
Emerging location intelligence platforms leverage interconnected sensors that take advantage of the decreasing cost of off-the-shelf LiDAR components and the ubiquity of smartphones, wearables, and wireless infrastructure (WiFi, Bluetooth, ultra-wideband) to track people and assets. LiDAR remains an ideal technology for these sensors because its high fidelity is maintained regardless of lighting conditions and, unlike video, maintains citizen privacy. The result is a platform that is autonomous and scalable, and operates in real time to deliver 3D models and 2D maps while incorporating location and positioning within a single solution. These sensors are small and can be deployed on top of a building’s existing infrastructure—not dissimilar to a WiFi router.
The advantage of this approach is that the platform is able to capture the data needed to create 3D models of the indoor space while also understanding the sensors’ own positions relative to each other. Each sensor captures point clouds of the same space from different perspectives to create a composite point cloud that is automatically converted into a single structured model. This solves the two critical roadblocks that industry has faced when trying to acquire indoor data: Maps no longer need to be created/updated by professional services teams, and the maps and positioning data are always integrated and updated in real time.
The sensors track electromagnetic signals from people and assets within the space. This approach respects citizen privacy by capturing a unique identifier rather than personal information. Algorithms eliminate redundant data (such as signals from computers or WiFi routers) when identifying humans within a space and model the traffic patterns and behaviors over time. The data includes the person’s or asset’s longitude and latitude coordinates, along with altitude—which, as more people live and work in high-rise buildings, is becoming increasingly necessary for emerging enhanced 911 requirements in the United States. The scalability and real-time nature of a platform-based approach results in a stream of data that can be used to drive a variety of applications, including wayfinding, evacuation planning, training and simulation scenarios, airport security, and more.
Integrating Indoor and Outdoor 3D
Accurate correlation in four dimensions will drive the framework for future information transfer and corroboration. Fixed objects at a point in time must be properly located for all of their properties to be associated. This work is more challenging than it might appear. Many objects look very similar, and various sensors have differing levels of resolution and accuracy—bleed-over of attribution and misattribution of properties is possible. The better the 3D base layer, or foundation, the more likely all scene elements will be properly defined. Once objects move within the scene, the correlation of observables, initial position, and the changes to it often allow inference of intent or purpose.
Connecting data from outside to inside to deliver a seamless experience has yet to be solved, although there is progress. By capturing indoor 3D quickly and in real time, the opportunity to integrate it with outdoor 3D models is now possible. We expect the integration of 3D and its related positioning data will soon be ubiquitous regardless of where a person is located. In areas where data providers can work together, the same approach used by Google to track traffic could allow for the establishment of routes from outdoor to indoor and vice versa to evolve rapidly. Companies creating the 3D data are defining the standards, and, as more data becomes available, accessing information can be as easy as “get.location” for software developers creating outdoor navigation apps. A centralized database with established formats, standards, and access protocols is recommended to ensure that analysts and developers work with the same datasets and that decisions and insights are derived from the same foundation, no matter where stakeholders are located.
3D Accessibility for Success
As it becomes easier to quickly and cost-effectively create and integrate indoor and outdoor 3D data, managing how that information is stored and accessed will be the next opportunity for the geospatial community. In order for 3D to be truly valuable, it must be easily—if not immediately—accessible for today’s devices. Ensuring 3D can be captured in real time will drive the need to deliver it quickly and across a wider variety of applications. A smart compression and standardization strategy is critical to the portability of the information. As the use of 3D by consumers increases, there will be more natural demand for ready access from user devices, which will help streamline and optimize applications (as it has for 2D mapping over the last decade).
Applying 3D to the real world, in real time, provides:
- Improved situational awareness to users from their own devices.
- Seamless wayfinding from outdoors to indoors.
- Exceptionally detailed and portable data for military/emergency planners and operators.
- Readily available data and web access for first responders and non-governmental organizations.
- Global GPS-denied navigation capability for mission critical systems (e.g., commercial flight avionics).
- A globally accurate positioning grid immediately available for analysis.
3D is ready to play a bigger role in how we experience the world. The manipulation of location in 3D should be as natural as controlling a video game console. As long as the GEOINT Community keeps in mind what has been learned from both 2D mobile mappers and gaming aficionados, the move into the Z-dimension should prove as easy as it is worthwhile. Moving forward, 3D data developers and users have an important role to play—to provide feedback on what “feels” natural, and what doesn’t. After all, that’s what reinserting the third dimension is all about.
Headline image: The Great Pyramid of Giza, from 3D compositing of DigitalGlobe imagery. Courtesy of Vricon.