In 2019, the Singapore Land Authority initiated another mapping project to detect changes since the last time the city was mapped in in 2012. The new 3D model of the city is reported to be more accurate than the previous one. The project required aerial mapping of the entire country and mobile street mapping of every public road. (Picture courtesy of Singapore Land Authority.)

Bentley Systems’ open and federated workflows enable data sources from various agencies to be integrated while keeping the software in its native file format. Above shows data from CityGML, GeoTiff, Shape and DEM files. In addition, data from spatial databases, reality meshes, CAD data, geodatabases and point clouds can be integrated. This data is now available for local businesses to use and introduce their own design elements to run checks and analysis with existing as-built context. (Picture and caption courtesy of Singapore Land Authority.)

You may be used to zipping around in a city with Google Earth or Street View in Maps. Without requesting a penny in return, Google seems to have driven every road on Earth with 360-degree cameras and uploaded all the data to the cloud, so without leaving the confines of your apartment, you can virtually visit the places the pandemic has denied you from accessing. But if you are an architect, an urban planner or a construction chief, Google’s views don’t cut it. There’s not enough detail—they are limited in perspective and distorted—and a little hard to navigate. Clearly, a more accurate and navigable 3D model is needed for professional use.

Enter Singapore, a city committed to be not just a smart city, but also the smartest city in Asia, and Bentley Systems, long a willing technology partner to firms and agencies with immense infrastructure … up to the scale of entire cities.

Bentley applications were key to creating an even better, geo-accurate 3D model than Singapore had created previously in 2012. With photogrammetry and scanning tools, drones took pictures of 160,000 buildings. From them, Bentley applications were able to stitch together an accurate and navigable model of the entire city/state of Singapore.

Singapore is nothing if not digitally driven. An IFR report listed Singapore as first in terms of robots per capita, passing South Korea in 2019. Singapore’s realistic-looking twin of the whole city/state may be first in terms of scale and accuracy of all smart cities.

The project gathered point clouds from LiDAR scans for every public street in Singapore. The sheer amount of data took up 25 terabytes of cloud storage.

With almost 6 million people crowded into the city, it must have been no small task to filter them out of the 3D model—something Google doesn’t bother with, choosing instead to blur their faces.

The Singapore 3D mapping had to literally rise above street view—this was very important for accurate modeling of the city’s many tall buildings and vertical expansions. Other islands with many tall buildings, notably Manhattan and Hong Kong, also must grow vertically, but none have it as a national mandate. Singapore counts its national area not just in terms of the ground but adds to it the area for each floor of its tall buildings. With only 500 acres within its borders (about the average size of a farm in the U.S.), more high rises make for a surefire area multiplier.

For most of the world’s cities, a 2D map may be sufficient for urban planning, but for a city that is obsessed with every usable area, including the floors of high rises, a 3D map is the only option.

The entire island was flown over to gather aerial images. Every public road was driven for ground-based images. Cadastral surveys ensured the precise position of buildings and structures, providing accuracy unobtainable from the popular Google Street View.

Singapore is somewhat guarded about the data for this project, not wishing to share it worldwide. The project was handled by a secure processing facility that housed the data on local cloud environment. The Singapore Land Authority (SLA) will make the data available locally, with the architecture, real estate, construction and logistics industries first in line, according to the July 2020 issue of SLA’s Land newsletter. For example, an architecture firm can see how a new building would look in the context of an area’s existing buildings.

A digitalization project of city scale required a software vendor that was comfortable with enormous models—and none are more comfortable than Bentley. The company counts among its users city and state agencies, including almost every U.S. state department of transportation. Projects like highway systems and railroad networks, national priority dams, city defining bridges, international airports and massive sports stadiums are routine business for Bentley. The robust portfolio of Bentley’s city and nation building product portfolio is all that was needed for a complex and detailed city model, including a photogrammetry app and mobile mapping technology, which in combination, would “securely process the voluminous data from multiple sources into a sustainable nationwide dynamic, 3D digital twin,” quoting a company press release.

“Bentley’s strength in interoperability enables the production of a 3D mesh, optimized for visualization or analysis with 3D software, 3D GIS, or web applications,” said Hui Ying Teo, senior principal surveyor at SLA.

Singapore Land Authority captured more than 160,000 high-resolution aerial images over a 41-day period and selected ContextCapture to process them into a 0.1-meter accurate nationwide 3D reality mesh. (Picture courtesy of Singapore Land Authority.)

The aerial images were taken over a 41-day period. Bentleys’ ContextCapture was able to stitch together the images. Given points of reference (from cadastral surveys) and LiDAR data, the resulting 3D mesh model is not only in full color but also geometrically accurate, varying plus or minus 0.1 meter across all of Singapore. Bentley’s Orbit 3DM was used to process 25 terabytes of LiDAR generated point clouds from ground-based vehicles.

“Bentley Orbit 3DM is an enabler to manage the large volume of point clouds and imagery, and bridge data sharing with other users with scalable cloud resources to support the long-term sustainable mapping program that SLA envisions,” added Teo.

SLA plans to update this recent model with additional aerial and ground-based data, rather than having to create a new 3D model in 6 to 7 years (the time frame between this model and the previous one).

Digital Twin Drives Smart National Development

SLA estimates that traditional methods of topographic surveys and aerial and street mobile mapping would have cost SGD 35 million (USD$26 million) and taken 2 years. Using modern methods and Bentley software let them get results with only SGD 6 million (USD$4.4 million) and took only 8 months. In addition, reusing the data is projected to avoid the 3-year cost of SGD 13 million (USD$9.6 million), resulting in significant savings.

More about the 3D Singapore Sandbox

SLA introduced the 3D Singapore Sandbox during the 2020 GeoWorks conference held in Singapore.

SLA initiated the 3D Sandbox project in March 2020, availing itself of software from Autodesk, Bentley Systems, Esri Singapore and Hexagon Geospatial, as well as “people movement intelligence” data processing by the San Francisco-based data analytics service LOTaDATA, which claims to help customers with “inferring people-presence, activity, and movement in the real world.”

SLA was able to create a 3D map, as it were, using aerial and street mobile imagery for a 3D city model throughout 2014 and 2015, but additional development and changing land use necessitated a new effort.

The 3D Singapore Sandbox project is part of the SLA’s “collaborative environment for its technology partners and developers. The city model and geospatial data will help partners to develop and test public and privately developed applications. The 3D models of 160,000 buildings will be made available to local businesses and agencies.”