Creating a nation-scale digital twin

Updating an aging nation-scale digital map

Singapore continues to thrive and develop, with new or revamped buildings and infrastructure constantly under construction, but the island nation has a finite amount of usable space. The Singapore Land Authority (SLA) carefully optimises and distributes land resources to ensure sustainable growth. “The limited land in Singapore is used for a wide range of activities and the effective use of land and space resources is paramount to the development of Singapore,” said Victor Khoo, Director, Survey and Geomatics, SLA.

As part of its mission, SLA began the National 3D Mapping Programme, which created a 3D map of the entire nation, first capturing aerial imagery in 2014, then street-level imagery in 2015. The imagery was combined and optimised by 2017, producing a detailed replica of the nation’s streets, buildings and terrain. In addition to helping officials plan continued development, the national government used the map for flood risk management and aviation safety.

However, Singapore is a dynamic country with sprouting developments and complex land use needs. As time passed, the 2014–15 snapshot of the nation became less accurate and less able to support urban planning. Additionally, SLA discovered significant data gaps in the national map. In 2019, SLA decided to conduct a second round of the National 3D Mapping Programme to fill in the gaps and illustrate changes made to the national environment.

Seeking improved detail and accuracy

As before, SLA planned to undertake both aerial and road-based image capture using vehicles with mounted cameras that could recreate the fine detail of buildings, and LiDAR devices that could accurately model the terrain. It would then process the images into a complete map of the nation. This time the intention was to use advanced technology to go beyond a simple 3D re-creation and create a true reality mesh that would provide a more detailed, accurate representation of the real world. Additionally, SLA wanted to be able to frequently and easily update the reality mesh with new data. “Our program will be a sustainable one with continuous maintenance to ensure data accuracy,” said Hui Ying Teo, Senior Principal Surveyor at SLA.

In addition to the accuracy requirements, SLA wanted to build the updated map in an open format, allowing data captured once to be seamlessly used by numerous government agencies that work with a variety of different applications. It also needed to easily handle the huge amount of data required to map the entire nation. SLA sought a solution that would meet all these requirements and provide an advanced 3D map of the nation that could be used for years to come.

Managing terabytes of data in an open format

Already familiar with Orbit 3DM, SLA determined the new generation of that application combined with ContextCapture could create the detailed and flexible reality mesh desired. “Other non-Bentley applications have been explored in the past, but none had the strength to process the scale of the project we had,” Teo said. Over 41 days, the team captured 162,625 images with 80% forward and 60% side overlaps, representing the entire 720 km² nation. ContextCapture’s processing workflow at the block-and-tile level improved project management and sped final editing needed to ensure the map accurately represented unusually shaped buildings. Additionally, they were able to use the applications offline, as the reality mesh was created in a secured processing facility to meet security requirements.

Once the reality mesh was completed, SLA used Orbit 3DM to share it with other stakeholders. The open application not only supports export into any format, it is also robust enough to manage huge volumes of data — the representation of the nation’s public roads alone totals 25 terabytes. Various governmental agencies can now view the map within familiar third-party software or Orbit 3DM’s viewer, which can be used with minimal technical knowledge.

Saving costs with continual map updates

By using applications that supported advanced aerial and street mobile mapping instead of traditional topographic surveys, SLA completed the map in eight months instead of two years and realised a cost saving of US$29 million. Additionally, the accessibility of the map improved by 60%, which will lower the time needed for urban planning. The completed nationwide reality mesh model is now a cornerstone of nine different government agencies and adds value to numerous projects.

The map, accurate to within a 10th of a centimetre, helps all stakeholders plan, design, construct and monitor infrastructure and has become a key enabler for developing Singapore into a smart, resilient and sustainable nation. With the much higher level of detail, Singapore agencies are using it to support more comprehensive initiatives, such as optimising 5G deployment, improving flight safety management, undertaking fluid dynamics and urban wind tunnel simulations, and optimising placement of surveillance services.

By using Orbit 3DM’s ability to manage huge amounts of data, SLA anticipates updating the existing map for many years. “Orbit 3DM is an enabler to manage the large volume of point clouds, imagery and bridge data and share it with other users by using scalable cloud resources to support the long-term sustainable mapping programme that SLA envisions beyond 3D and into the fourth dimension of time,” Teo said.

By being able to frequently refresh the existing map rather than completely recreate it every three years, SLA anticipates saving US$13 million each year while continuing to use it for optimal site planning.  SLA is continuing to explore potential new uses for the nation-scale map to enhance Singapore’s economic and social development, eg, integrating machine learning into the reality mesh for enhanced analysis of transportation models.

Image credit: iStock.com/ lena_serditova