Between the Poles

Estonia developing a national digital twin to increase construction productivity

Estonia is well known as one of the most digital countries in Europe.  At the GEOBIM 2020 conference this morning Christopher-Robin Raitviir, of the Estonian Department of Digital Construction in the Ministry of Economic Affairs and Communications, described how Estonia is adopting a BIM-based building permit process which is contributing to the development of a  national 3D digital twin.  The business objective behind these initiatives is to improve construction productivity by a factor of three.

The objective of the digital permitting process is to speed up permitting and raise the level of compliance enforced by the process. The digital permitting application accepts BIM models in IFC format and then applies rule-based and algorithm-based checks to ensure compliance with national standards.  For example, a standard for door size can be checked automatically and any doors that are not compliant are highlighted in red and identified and visualized in the 3D model making it easy to identify and correct problematic doors.  After permitting approval building designs are deposited in the National Building Registry.  This is interesting because it addresses a widespread problem; what to do with designs after the permit has been issued.  The designs have been submitted as part of a building permit application and have been publicly available for a period of time.  They contain important information that can be very useful for emergency responders and for other purposes.  The designs are often left to molder away in the basements of municipal governments because of uncertainty about ownership - even though the designs have been publicly available during the permitting process, access after permitting is often restricted because the information is considered to be proprietary.  Currently the models in the National Building Registry are designs, but the ultimate goal is for it to contain accurate as-builts.

The Estonian National Digital Twin was initiated in 2018 and includes buildings, roads and other transportation infrastructure such as rail and airports, and, remarkably, underground utility networks.  This data is publicly available and buildings, transportation links and utility networks can be selected for download in CityGML and other formats.

This is a remarkable initiative on a national scale and is similar in many ways to Singapore which has had a sustained effort to improve the speed and comprehensiveness of the permitting process by requiring digital BIM models.

Above and underground infrastructure comprise Rotterdam's open digital twin

The City of Rotterdam is developing a digital twin of the city.  3D vector and image data of the above- and underground infrastructure, zoning regulations and city by-laws, data streams from real-time sensors, applications and simulations will comprise Rotterdam's digital twin. It is envisaged that the digital twin will enable the city to assess the impact of proposed changes in bylaws and regulations by simulating them before they are implemented. Rotterdam, and in fact the whole of the Netherlands have adopted transparency and open data as a pillar of government and are basing 3D models on open national and international standards, intend for the 3D data to be open and accessible to the public, and intend to use open source software wherever possible.

Many cities have 3D models of their aboveground infrastructure; buildings, transportation networks, parks, and other infrastructure typically captured in 2D or 3D imagery from overflights or in 2D GIS maps.  Imagery is not very smart and extracting 3D vector objects such as building footprints, heights, profiles and transportation networks requires sophisticated post-processing.  Even more importantly, underground infrastructure is usually neglected, even though it is recognized that water and wastewater, energy (gas, electric power, and district heating),  and communications (fibre and copper) networks provide the life blood of the city.

Rotterdam has been developing its 3D model, which underpins the digital twin, for several years.  Currently it contains street furniture such as street lights, trees and underground parts of street furniture, a large scale basemap, a complete map of underground networks of cables and pipes, buildings, and non-physical content such as zoning regulations and  bylaws.  The 3D model is kept in sync with data repositories maintained by city departments and others.  In the future it will also include real-time data streams from a variety of sensors. 

The 3D model is open, based on open standards and many of the applications that have been developed for it are open source projects. For example, visualizations rely on Cesium, an open source initiative which has just announced support for visualizing underground utilities, mines, and geological layers. The rationale for the data being based on open standards is that an standards-based environment creates an open data commons where developers can devote their time to developing innovative applications rather than spending time converting data from one format to another.

The Dutch 3D standard is an integration of CityGML Application Domain Extension (ADE) and the existing national IMGeo standard for large-scale representations of roads, water bodies, and land use. The Dutch Kadaster, Geonovum, the Netherlands Geodetic Commission and the Dutch Ministry of Infrastructure and Environment initiated a "3D Pilot NL" to develop a national 3D standard.  Existing intnernational standards including CityGML(Open Geospatial Consortium), IFC (buildingSMART), and others were investigated. The findings of the pilot  resulted in a national 3D standard implemented as a CityGML Application Domain Extension (ADE) that integrates the latest version of the existing Dutch national standard IMGeo 2.0. 

Current applications (virtualcitySYSTEMS, ESRI, IMAGEM and Unity) use the 3D model as input for analytics and visualization. One can imagine many use cases for the data in the Rotterdam 3D model. Among these three use cases have been identified as low-hanging fruit; climate adaptation, viewsheds and energy performance of buildings. 

Much of Rotterdam is below sea level and managing rain water accumulation during heavy rainfalls is a major issue. which affects building basements, subway entrances, low-lying roads, electrical transformers, and other infrastructure.  To address this problem, the concept of "water squares" has been developed and the 3D model used to determine the optimum location for these..  Water squares are open areas where water can safely be allowed to accumulate during heavy rainfalls to avoid overstressing storm sewers. 

In another application of the 3D model data, building data can be filtered and streamed to energy performance applications that provide insights into how buildings perform.

Open standards and open data also enables the sharing of data among cities in the Netherlands.  The Dutch Kadaster has developed an open 3D map of the whole of the Netherlands, with scales ranging from 1:500 to 1:25,000 in collaboration with academic, private sector, and public sector partners.  All data in the 3D map is built on the basis of open standards.  The map was created by combining existing 2D data with the Elevation Model of the Netherlands resulting from a LiDAR survey which covers the whole of the Netherlands with a resolution of eight elevation points per square meter.

Open data and open standards also relate to a critical issue that is of growing importance. Many governments have adopted transparency and open data as a cornerstone of democratic policy.  But often digital twins for smart cities are being developed by private-sector firms or by public-private initiatives.  As William Davies points out in Nervous States, "the IT infrastructure used to build e-geovernment projects or "smart cities" comes from the commercial sector, and the companies that provide it often retain control over the data that results." 

For example, recently a joint Alphabet/Google/Sidewalk Labs and City of Toronto initiative to develop a smart city on Toronto's waterfront foundered and was cancelled.  One of the reasons, it is argued, is that a technology giant such as Google (or Facebook, Amazon, ...) is not a vendor in the traditional sense, but a "surveillance capitalist" that collects and analyzes behavioural data captured from users for its own economic imperatives.  In the context of a smart city the opportunities for data collection from citizens and from systems such as water and wastewater, transit, waste removal, traffic management, and others are immense.  If this data is appropriated and becomes proprietary, it puts a private entity in a position to supplant democratic government, which doesn't have access to this huge volume of data and the analytics which extracts information from it.  Relying on open standards and ensuring that data is open and accessible to the public are essential to avoid citizens' and municipal data becoming proprietary.  Furthermore, using open source software and, I would add, open training data for machine learning applications helps avoid the problem of the analytics also becoming proprietary.

Based on a Webinar by Christian Veldhuis, Advisor City of Rotterdam, Rick Klooster, CCO, FUTURE INSIGHT and Antje Kunze, Director Sales & Marketing, virtualcitySYSTEMS.

The future is fewer maps

At the first Longitude meetup in the U.K., organized by Ben Flanagan, Kenneth Field, and Chris Wesson, Ed Parsons presented his perspective on one of the directions to expect from geospatial in the future.  It was a bit counterintuitive, for consumers more geospatial data and technology, but fewer maps.

To start with he reminded us that the public has never had as much access to maps as we do now thanks primarily to Google Maps - either as standalone or embedded in other apps. What he foresees is that in future we will be seeing fewer maps or no maps at all.

The maps we have used up to now are large scale maps presenting a large scale view - probably very useful to Winston Churchill and people planning things on a large scale. But in our day to day life, we don't need maps at this scale. Instead we need to know what we will encounter in the next 15 minutes, whether that be the next exit off the highway or the next grocery store. There will be more interaction between us and the world around us - for example showing which of the nearby bars or coffee shops are very busy or less busy.

Ed gave a practical example that anyone who uses a smart phone to navigate in a place they have not been before can relate to. When you come out of a tube station or any GNSS-denied place you are always faced with a quandary - how to orient yourself to know whether to walk left or right. I mentally flip a coin and walk until my smart phone tells me if I am headed in the direction I want to go. If I am in a part of the city where GNSS is unreliable I walk to the end of the block and the street number or name usually will tell me if I am headed in the right direction or that I have to retrace my steps in the other direction.  In the future your navigation app won't rely solely on GNSS signals but will recognize buildings and give you the right direction. In fact this could all happen through an audible interface so that you needn't look at a map at all, "turn right as you leave exit 3 from the tube station".  Of course this will require a lot more spatial data about the city you are in and will require some serious geoanalytics.

What this suggests is that as consumers we will use more geospatial data and analytics, but it will become part of the technology stack and most of us won't realize how much geospatial technology we are using.

Bentley's digital twin capabilities for cities includes underground infrastructure

Proposed West Link tunnel under central Gothenburg

More and more cities are recognizing the tremendous benefits that a digital twin provides to its planners, private architects and engineers striving for compliance with city ordinances and by-laws, and for communicating with the public.  I remember several years ago when the Prime Minister of Singapore in his annual National Day address included a 3D visualization of what the proposed Marina Bay towers would look like - perhaps the first time 3D visualization has been used to inform citizens on a national scale.  At the Year in Infrastructure 2018 in London, Bentley announced the acquisition of Agency9, a Swedish company that provides web and mobile solutions for 3D visualization of geographic information and maps for collaboration and communication with its product CityPlanner (now OpenPlanner).  With this acquistion, Bentley augments its competitive positioning relative to CityZenith (SmartWorldPro), ESRI (CityEngine), and Autodesk (Infraworks) and open source projects such as mago3D in offering a 3D visualization and analytics platform for the massive data sources required for cities including underground infrastructure.

The requirements of a comprehensive platform for city visualization and analysis for planners, architects and engineers, and the public are the ability to support OGC (KML, CityGML and others), BuildingSmart (IFC), defacto (Shape, DWG, DGN, Revit and others) and other standards such as GeoJSON making it possible to integrate GIS, CAD, and BIM (vertical and horizontal) data, content creating and editing tools, for example, to allow buildings to be extruded from footprints and rooves and facades customized, support for open data sources such as OpenStreetMap, ability to integrate satellite, aerial and ground photo imagery and LiDAR point clouds,  support for both above- and below-ground infrastructure, gaming engine support (such as Unity or Unreal Engine) for interactive 3D visualization, cloud processing for managing, processing and querying massive amounts of data, web and mobile access, and APIs for analytics.

From a technical perspective the acquisition made a lot sense because Bentley's ContextCapture (from the acquisition of Acute3D several years ago), which creates 3D meshes for both inside and outside buildings from aerial and ground photo images and LiDAR point clouds, has been used by Agency9 for many years as the basis for its visualization capabilities.

Proposed West Link tunnel under Gothenburg

The West Link (Swedish: Västlänken) is a railway tunnel under construction under central Gothenburg. As an example of its support for underground infrastructure the City of Gothenburg used CityPlanner to publish a 3D visualization of the planned tunnel going below the entire city which showed stations with points of interest with images in a slideshow. During the first 24 hours that the project was published on its web site, it had over 10,000 unique visitors.

Progress in the UK toward a digital twin of the underground

Productivity Growth – Output per worker (1997=100) Source: ONS

Construction productivity is a challenge in the world's advanced economies. Globally, labour productivity growth in construction has averaged only one percent a year over the past two decades, compared with growth of 2.8 percent for the total world economy. In the Unites States labour productivity in construction has actually declined.  A recent report from the UK Government Infrastructure and Projects Authority, "Transforming Infrastructure Performance", provides evidence for stagnant productivity in the UK construction industry and outlines the government's plan to address this problem including development of a digital twin.

A recent McKinsey study revealed that construction is one of the least digitized industries in the world. McKinsey Global Institute’s Industry Digitization Index combines 27 indicators to measure the digital assets, digital usage, and digital workers in each sector. Construction is among the least digitized sectors in the world, at a level comparable to agriculture and hunting. According to the McKinsey index construction comes second to last in the United States and last in Europe.  Wipro, Booz Allen Hamilton, Aconex,  Jacobs, McKinsey, Hexagon, Trimble, Bentley, Autodesk, and many others believe that the construction industry is poised for a digitization revolution and are investing heavily in technology to enable this to happen including geospatial, BIM and full lifecycle BIM, machine learning, drones, mixed reality, big data, analytics, automation and hyperautomation.  There are an increasing number of technology startups in the construction industry.  There are also promising signs of interest in underground detection and information sharing.

The Transforming Infrastructure Performance report reiterates the widely held view that the UK construction supply chain "as a whole is under‑achieving. It has low profitability and invests too little in capital, research and development and training.” Transforming Infrastructure Performance (TIP) is the UK government’s plan to increase the effectiveness of investment in infrastructure including transport and energy networks, and social infrastructure (schools and hospitals) by improving productivity in the way assets are designed, built and operated.

The Digital Built Britain program is a key part of the government’s plan to develop the UK's digital data capabilities which included developing a ‘digital twin’. In the UK it is estimated that new assets add less than 0.5% each year to the value of existing infrastructure, which highlights the importance of improving the performance of existing assets. It is expected that a national digital twin of the real world estate will help inform government decisions to maximize full lifecycle performance and benefits of assets.

A digital twin includes above and below ground assets. Key dimensions of a digital twin are location and condition of each asset. The UK has two research projects focused on the underground part of a national digital twin. Mapping the Underworld aimed at developing technology to locate in 3D all buried utility assets without excavation.  One of the key outcomes of this project was estimated direct, indirect and social costs of utility strikes during excavation by researchers at the University of Birmingham.  The second phase Assessing the Underworld focused on geophysical tools to assess the condition of buried pipelines and cables, and of the ground in which they are buried, and of the surface transport infrastructures beneath which they are buried.

Combining above and below-ground information into one national single data model/data exchange framework will provide a foundation for the development of smart cities. Project Iceberg is an exploratory project undertaken by the British Geological Survey, Ordnance Survey and the Future Cities Catapult to investigate ways to integrate data and services relating to the underground with other city data. To date two reports Market Research into the Current State of Play and Global Case Studies and Defining the problem space for an integrated data operating system above and below ground have been published.The medium term objective is to take these concepts forward with project partners to develop new digital data demonstrator projects.

Towards a digital twin of infrastructure: on the cusp of efficient reality capture of underground utilities

Public safety is at risk whenever the location of underground infrastructure is unknown or inaccurately known. For example, in Belgium workers on a construction project rupture a high pressure gas line they did not know was in the vicinity and the resulting explosion results in fatalities and injuries.   Or a house explodes when a gas line penetrates a sewer line during trenchless horizontal drilling and an attempt is made to clear the resulting obstruction with an augur.  Or it requires hours to identify the source of an explosion in a residential neighbourhood in San Bruno, California which caused fatalities. Similar events have occurred in Sun Prairie, Wisconsin, Allentown, Pennsylvania, Bellingham, Washington, El Paso, Texas, Saudi Arabia and a number of other locations.

Damages and delays connected with underground utilities during construction represent at a minimum a $50 billion drag on the U.S. economy.  For example, a project in Sydney, Australia to extend a light rail transit system encountered many unknown underground utilities during construction which delayed the project by one and half years. In the U.S., a leading cause of highway construction delays is missing or inaccurate information about the location of underground utilities. To address the risk of liabilities associated with unknown or inaccurately located underground utilities, contractors regularly increase bid costs by a minimum of 10-30%. 

A digital twin of the underground would eliminate or ameliorate these issues.  Creating a digital twin requires addressing challenges.  First of all, considerable effort goes into detecting the location of underground utilities in preparation for and during construction on virtually all projects, but this data is rarely shared.  If a way could be found to share this data, not only would contractors not have to rediscover the location of underground utilities over and over again, but over time a 3D map of underground infrastructure would emerge at no significant additional cost to construction contractors.  Secondly, a technological advance is required that enables the efficient, safe and rapid capture and processing of remote sensed data to create accurate 3D maps of the underground.  Recently there have been advances in addressing both of these issues that provide grounds for optimism that we are close to being able to cost-effectively create a digital twin of underground infrastructure. 

Putting this in the context of smart cities, where data in the form of a national digital twin becomes just as important as physical assets, it is essential that a national digital twin include above and below ground assets. Fundamental to this process is rethinking value, not just the value of a finished building or infrastructure asset, but over the entire lifecycle of an infrastructure asset. Furthermore, managing this data cost-effectively requires finding ways to integrate data and services relating to the underground with other city data.

Why a digital twin of underground infrastructure is important

The impact of poor information about the location of underground utilities is exemplified in the Sydney Light Rail Project.  This is a $2.1 billion PPP project for a 12 km of light rail extension in Sydney, Australia to be completed by 2019. In preparation for construction a year was allocated for identifying potentially conflicting utilities in the proposed right of way. About 500 subsurface utilities were identified and scheduled for relocation. During construction a further 400 unmapped utility services were encountered. An economic impact study by an independent consulting firm estimated that the project could have been completed at least one and a half years sooner if a complete and reliable 3D map of underground infrastructure had been available at the project planning stage.

In the U.S. underground utilities are a leading causes of highway construction delays.  Almost all construction projects for buildings or infrastructure contain a design constraint: existing utilities strung overhead on visible structures or hidden underground.  The traditional strategy for the highway designer has been to ignore utilities during design. Prior to construction utilities in the areas were requested to locate any of their underground infrastructure in the proposed right of way. Utilities are routinely relocated, often at great expense, often unnecessarily and often negatively impacting budget and schedule of the construction project. A more cost-effective alternative “design-to-accommodate” is to locate utilities first and then design the highway to avoid them, but that requires reliable information about where the utilities are located. This is rarely available. Between the two alternatives of relocating utilities and designing the highway to avoid utilities whose location is poorly known designers try to find a workable compromise that meets the highway construction scope and mission, while minimizing impacts to utility facilities. If successful, this can result in substantial savings in utility relocation costs and impacts, as well as overall savings to the project budget and timeline.  A survey has found that state departments of transportation (DOTs) overwhelmingly would like to get utilities involved as early as possible to reduce project risk by determining what utility infrastructure is present in the project area and where it is located. The study found that there is a general consensus that this would help make better decisions regarding relocation versus design-to-accommodate. When comprehensive and accurate utility location data is available, better decisions can be made, and the risk of unforeseen problems due to utilities emerging during the construction phase are reduced.

According to the Common Ground Alliance (CGA), in the United States an underground utility line is hit on average every 60 seconds, that is 390,366 hits in 2016. In addition there is always the risk of injuries and fatalities. I have estimated that the total cost to the national economy is over $50 billion. The problem is a location one; we simply don't know where our underground infrastructure is.  In most municipalities in North America underground utility lines have been put in the ground not according to plan but wherever it has been easiest and cheapest to build them. As a result 2D as-builts of underground infrastructure are notoriously unreliable. Neither the local municipality or the utility network owner know the location of underground utilities.

Construction bids routinely include an added cost, which can amount to 10%-30% of the bid cost, to cover the risk associated with underground utilities. In spite of the effort in locating underground infrastructure that is part of virtually every construction project the information is rarely shared and as a result the location of the same underground utilities is often captured over and over again for different construction projects.

As a result locating underground utilities is an essential part of every construction project involving excavation and has become a major industry in the U.S. The Common Ground Alliance (CGA) which is an association of companies and practitioners in North America reports 1700 members in the U.S. and Canada. The last CGA annual conference and exhibition attracted 800 attendees and over a hundred exhibitors which gives a feeling for the size of the locate sector.

In the UK about 4 million excavations are carried out every year on the UK road network to install or repair buried utility pipes and cables. Not knowing the location of buried assets causes practical problems that increase costs and delay projects, but more importantly, it increases the risk of injury for utility owners, contractors and road users. The problems associated with inaccurate location of buried pipes and cables are serious and are rapidly worsening due to the increasing density of underground infrastructure in major urban areas. Research at the University of Birmingham has determined the direct costs of utility strikes in the U.K.

Facility	Avg cost per strike
Electricity	£ 970
Gas	£485
Telecom	£400
Fibre-optic	£2800
Water	£300-980

The researchers have found that the true costs, including indirect and social costs, associated with utility strikes is 29 times the direct cost.

In the U.S. the Common Ground Alliance reports that there were 390,366 documented underground utility hits in 2016. The average costs of the damage for each utility hit were calculated for different types of underground infrastructure.

Facility		Avg cost per damage 2016
Natural gas		$5,914.00
Telecom		$3,022.00
Electric		$4,905.00
Cable TV		$2,190.00
Water		$3,003.00
Sewer		$5,163.00
Liquid pipeline		$7,711.00
Steam		$1,800.00
Average		$4,021.00

At an average cost of $4000 per hit, the CGA estimates that the direct cost to the U.S. Economy was $1.5 billion. The CGA emphasizes that this is a conservative minimum estimate and does not include indirect and social costs such as property damage, evacuations, traffic disruption, road closures, resulting lost business, environmental impacts, lawsuits, injuries, and fatalities. If the ratio of total to direct cost determined by the researchers at the University of Birmingham is applicable to the US, it would mean that the total estimated impact of unknown or poorly located underground infrastructure on the U.S. economy is over $50 billion.

Return-on-investment of accurately mapping underground utilities

University research has put a dollar figure on the benefits of accurate location data for underground utilities.

According to a USDOT sponsored survey conducted by Purdue University in 1999, two broad categories of savings emerged from using SUE — quantifiable and qualitative savings. The Purdue study quantified a total of US$4.62 in avoided costs for every US$1.00 spent on SUE. Although qualitative savings (for example, avoided impacts on nearby homes and businesses) were not directly measurable, the researchers believed those savings were significant, and arguably many times more valuable than the quantifiable savings.

In 2004 in Canada, the Ontario Sewer and Watermain Contractors Association commissioned the University of Toronto to investigate the practice of using SUE on large infrastructure projects in Ontario. This study, entitled Subsurface Utility Engineering in Ontario: Challenges and Opportunities, determined that the average rate of return for each dollar spent on SUE services on those projects that could be quantified was $3.41.

In 2007, the Pennsylvania Department of Transportation commissioned Pennsylvania State University to study the savings on Pennsylvania highway projects that used SUE in accordance with the mapping provisions of the American standard. In their unpublished report, Subsurface Utility Engineering Manual, Pennsylvania State University found a return on investment of US$21.00 saved for every US$1.00 spent for SUE when elevating the quality level of subsurface utility information using SUE.

In 2010 researchers at the University of Toronto took an in-depth look at nine large municipal and highway reconstruction projects that developed an enhanced depiction of buried utilities. Based on this analysis, a cost model was proposed that takes into account both tangible and intangible benefits. All projects showed a positive return-on-investment (ROI) that ranged from $2.05 to $6.59 for every dollar spent on improving underground utility location data.

In Europe an economic analysis of a Milan pilot project to map all underground infrastructure using ground penetrating radar for the Expo Milan 2015 site was carried out. The study found a return on investment of €16 for every euro invested in improving the reliability information of underground infrastructure.

In the U.S. the Federal Highway Authority (FHWA) cites studies that have shown that the cost of detecting underground utilities prior to construction typically costs less than 0.5 percent of the total project construction cost, saves more than $4 for every $1 spent, and can reduce project delivery time by as much as 20 percent.

Growing private investment in infrastructure drives technology investment in construction

The need for investment in infrastructure has continued to grow. It has been estimated that some $57 trillion will be required over the next decade and a half, or an estimated $90 trillion if we want to ensure that this infrastructure is sustainable. Increasingly infrastructure projects require private investment because governments simply have less and less money for capital projects. To fill the shortfall requires private capital. Private capital in pension funds, insurance companies, and sovereign investment funds amounts to over $50 trillion. Beginning in 2004 some of this money began to find its way into infrastructure investment and by 2007 private investment in infrastructure had reached $ 40 billion annually before the world financial crisis led to a cutback. Since then private investment in infrastructure has rebounded reaching nearly $80 billion in 2016.

Until recently construction has been at the bottom of the list with hunting and agriculture among industries with respect to investment in research and development.  There are signs that this is changing. In the last few years the number of startups focussed on technological innovation in construction has risen dramatically. Partly this can be ascribed to technological innovation that makes technologies such as reality capture, building information modeling (BIM), and augmented reality more accessible, but this is amplified by growing private investment in infrastructure. Private investment, unlike government investment which aims at social benefits, seeks financial returns. Whether the source of the funds is a pension fund, an insurance company, or a sovereign wealth fund, investments are expected to show a high likelihood of generating a significant financial return. This in turn drives productivity because increasing productivity of construction and operations and maintenance increases the return on investment.

Accelerating R&D into underground detection and modeling

There are indications of increased research and development initiatives related to underground infrastructure. In the U.S. for the first time the Common Ground Alliance (CGA), which represents the utility locate industry in North America, has just released a Technology Report, which describes the current state of underground detection technology including the latest developments and the gaps which CGA has identified in current hardware and software.

In the UK there have been quite a few research initiatives relating to underground infrastructure.  Project Iceberg, a collaborative project between Ordnance Survey (OS), British Geological Survey and Future Cities Catapult, has been created to better capture, collect and share data about underground assets and geological conditions and to find a way of sharing all this information among utility and energy companies, the transport sector, street works planners, building developers, and the construction industry.  The Mayor of London and the Smart London Infrastructure Network, which is composed of utilities responsible for water, energy, telecommunications and waste management in London, sponsored a challenge with the objective of developing digital technologies to accurately locate underground assets, including pipes, cables and joints, and determine their condition. The goal was to improve safety, and to reduce operational downtime, cost and environmental impact, and disruption. A national 10 year academic research program Mapping the Underworld focussed on remote-sensing for locating sub-surface utilities has just completed and been replaced by a follow-on 10 year program Assessing the Underworld. The BIM for the Subsurface (2015-2017) project, funded by Innovate UK, aimed to address issues such as project delays due to unforeseen ground conditions by applying the BIM process directly to ground investigation and subsurface infrastructure design.

Across Europe the SUB-URBAN COST Action (2013-17) aimed at developing relationships between experts who develop urban subsurface knowledge and urban decision makers, practitioners and the wider research community. The Action established a European network across 30 countries of Geological Surveys, Cities and Research Partners.

IDS GeoRadar Stream C

There is evidence of growing investment by major vendors in underground detection technology.  Hexagon, with revenue of $4 billion annually, acquired IDS (Ingegneria dei Sistemi) GeoRadar, an Italian company that has developed a towed multi-channel ground penetrating radar array that can collect data on underground infrastructure at faster speeds than has been possible heretofore. Since the acquisition Leica continues to release important enhancements to the software used to create 3D models of underground from the GPR scans and other information.  Furthermore a major software company, Bentley Geosystems with about $1 billion in annual revenue, has developed software specifically for managing location information about underground infrastructure. Bentley Subsurface Utility Engineering (SUE) enables users to develop 3D subsurface utility models from 2D models, resolve subsurface utility clashes, and model, analyze, and design subsurface infrastructure networks.

Recent advances in detecting and sharing location of underground utilities

Currently standard practice requires walking a lawn-mower like GPR array or hand-held EM wand along a road or right of way to detect underground utilities.  This is not only time-consuming and expensive, in traffic it is extremely dangerous.  Recently technical advances in detection were reported that bring rapid, cost-effective and safe collection of location information about underground infrastructure with no boots on the pavement closer to reality. T2 Utility Engineers, based in Whitby, Ontario have reported commercially using a multi-channel ground penetrating radar (GPR) array towed routinely at 10-12 km/hr to capture subsurface data.

Siteco rig with Sensors and Software GPR array and Faro LiDAR scanner

In a separate initiative a successful proof of concept by DGT Associates has been reported in Mississauga, Ontario in which data collected by a rig combining mobile laser scanning and GPR arrays collected data simultaneously above and below ground towed at roadway speeds of 80 to 90 km/hr.  In both cases the data was collected in soil conditions that were less than optimum for GPR. In both cases no boots were required on the pavement which addresses a serious safety issue with traditional GPR and electromagnetic (EM) detection.  However, in both cases significant time was required to post-process the collected scans, combine this information with as-builts, EM scans and other information to create 3D models. Much of the post-processing work remains manual. A rough estimate is that the post-processing requires two weeks per ten kilometers.  Nevertheless, that data can be collected safely at or near roadway speeds overcomes an important hurdle in developing 3D models of underground infrastructure.  In addition improvements in software are helping to speed up the post-processing required to create 3D models.

A significant cost for virtually every construction project is detecting underground utilities prior to and during construction.  This information is rarely shared so that the information about underground utilities collected during construction is effectively lost has to be collected over and over again. There are several initiatives underway to develop ways to share information about underground infrastructure that is captured during construction.

City of Chicago, Univ of Illinois Real-Time and Automated Monitoring and Control Lab (RAAMAC) and CityZenith

The City of Chicago has launched a pilot to deploy a platform for collecting data and creating and sharing a 3D map of underground. It is based on new technology developed by University of Illinois at Urbana-Champaign’s Real-Time and Automated Monitoring and Control Lab (RAAMAC) and Chicago start-up CityZenith. During excavation a dozen or more pictures are captured with an inexpensive digital camera. RAAMAC's software uses the photos to create a 3D digital model of the underground infrastructure. These models can be securely shared between the City of Chicago and construction contractors to improve project planning and limiting accidents. The advantage of this approach to data collection is that it does not interfere with construction and does not add any significant cost.  In a similar vein Bentley Systems has experimented with a system that equips excavation equipment with four inexpensive digital cameras that are used to collect images of underground infrastructure encountered during excavation. These images can be used to create an accurate georeferenced 3D map of the utilities encountered during excavation. A research project involving Costain and Bentley Systems has found that digital photography during excavation using a consumer grade smartphone can be used to create a 3D model of comparable accuracy to a laser scan survey and at much lower cost.

Recent advances in standards for locating underground utilities

Recently there have been developments that reflect improvements in underground remote sensing technology. Standards for reporting the reliability of the locational information about underground utilities have been in place for decades. In the U.S. the 2003 ASCE 38-02 which has been used for classifying location information about underground infrastructure according to its estimated reliability, is widely seen as being out of date. In France the 2012 presidential decree defines three explicit levels of cartographic accuracy for underground structures; A - less than 40 centimeters, B - 40 centimeters to 1.5 meters, and C - greater than 1.5 meters. In the UK the 2014 Publicly Available Specification (PAS) 128, developed under the auspices of the British Standards Institution (BSI) and sponsored by the Institution of Civil Engineers (ICE) and others, not only includes the A,B, C, D quality levels of the U.S. standard, but extends it with explicit precision levels B1 to B3. A process to revise PAS 128 to reflect newer technical developments has just been initiated.

Three organizations have sponsored a concept development study by the Open Geospatial Consortium (OGC) for underground information with the main outcome being an Engineering Report. The sponsors of the OGC Underground Concept Development Study were the Fund for the City of New York - Center for Geospatial Innovation, the Singapore Land Authority, and the Ordnance Survey (UK).  The Engineering Report documents the progress made by the OGC and its members to build a complete situation assessment and develop a conceptual framework for action to improve underground infrastructure data interoperability. The report also identifies the most important steps to be taken next in order to develop the necessary data standards.

This led to the development of draft versions of a Model for Underground Data Definition and Interchange ("MUDDI Data Model") and an RoI (return on investment) model Cost Benefit Assessment of Subterranean Information Management. MUDDI is being developed as an interchange and integration model to support a range of critical underground infrastructure and environment information that is currently held in a wide and disparate variety of forms. To review and assess MUDDI through actual implementations, the OGC has issued a call for participation for a MUDDI Workshop to be held July of this year in New York City.

Municipal and regional mapping of underground infrastructure

Some cities and regions have realized the value of knowing where their underground infrastructure is and have required better information about the location of underground utilities.  In most cases this involves legislation requiring and implementation of a system for sharing information about underground utilities between utility and telecom companies and construction contractors.

Many years ago Tokyo developed the mainframe-based Road Administration Information Center (ROADIC) system which was deployed first in Tokyo and then in most major cities in Japan. The ROADIC system provides information about the location of underground infrastructure including telecommunications and utilities.

Sarajevo, in Bosnia, has recorded the location of all utility and telecommunications infrastructure operating in the city on paper maps for over 50 years.  The building permitting process in Sarajevo required contractors working within the city to supply coordinates of all underground infrastructure to the permitting office where they were transcribed onto paper maps.  A decade ago Sarajevo converted these maps to digital format (Oracle Spatial).

Calgary, Alberta has had the JUMP (Joint Utility Mapping Project) for many years.  A city by-law requires that anyone placing cables or pipes undergound within city limits provide 2D maps in electronic form showing the location of the infrastructure.  The maps only show location and type of infrastructure (water, gas, electric power, telecom) and owner.  Anyone excavating in Calgary can access this database.  Edmonton, Alberta also has a shared facilities mapping database. The Instituto de Información Territorial del Estado de Jalisco developed an integrated infrastructure database for the State of Jalisco.

A project of the City of Sao Paulo called GeoCONVIAS which integrates data from 20 to 30 utilities operating in the city of Sao Paulo, which is one of the World's largest cities. The most important objectives  of the ConVIAS project are to organize underground infrastructure, prevent accidents during excavation, reduce inconvenience to the public, and reduce the costs of maintaining underground infrastucture. The utilities in Sao Paulo are not asked to provide detailed information about their underground facilities, just "a simple line" showing the location of their facilities.

Rio de Janeiro has a similar project GeoVias funded by the government of the City of Rio de Janeiro and four utilties.  One of the immediate motivations for this project was "man hole explosions" apparently resulting from the close proximity of underground electrical and gas facilities in Rio.  The objectives of the project in Rio are somewhat different from Sao Paulo, apparently because the network of underground facilities isn Rio is absolutely unique, certainly in Brazil, but maybe even in the world.  The objectives of the Rio project were to reduce the risk of accidents, eliminate hazardous interaction between different networks, for example, electric power and gas and provide information about the location of underground networks for construction projects. It was hoped that the project would also speed up the process of reviewing construction permit applications, which currently can take a year to complete, and emergency repairs to underground infrastructure.

Penang, Malaysia developed an interesting approach for mapping and maintaining a database of underground facilities.  Sutra D'Bank (Penang State Government Subterranean Data Bankis a database that is maintained by a joint venture company Equarater (Penang) Sdn Bhd (EPSB) .  Sutra D' Bank's customers are utilities or any other party undertaking excavations in areas under the jurisdiction of the local government.  The operator of Sutra D'Bank does two things.  They will provide the location of underground facilities in the planned excavation area, using the Sutra D'Bank database of undergound facilities supplemented by an on-site survey using a variety of technologies.  This is similar to the service that many utilities and telcos provide,  but identifies all underground facilities, not just those of one utility or telco.  In addition, the operator of Sutra D'Bank conducts an onsite as-built survey to record the 3D position and all relevant attributes of the new installation which they then upload to the Sutra D' Bank database.  This is unique and addresses a common problem - as-designed vs. as-constructed as-builts - that many utilities and telecoms face who rely on their construction contractors for as-builts.

The remarkable Sydney Down Under project brings together utility infrastructure including water, wastewater, telecommunications, electric power, and transportation infrastructure including rail and subways, and roads and highways, together with buildings (above and below ground) including interior spaces and occupants. It was developed by the Emergency Information Coordination Unit of the New South Wales government. This project involves 200 organizations including local and state governments, utilities and the telecommunications company.  The project is focussed on emergency and disaster management. Since the information is intended for emergency response applications, its availability is restricted.

A major step forward in mapping underground infrastructure was a pilot project carried out using ground penetrating radar (GPR) on the site of the Expo Milano 2015 event in Milan. The total project area is about 230 000 square meters. All underground infrastructure including electric power, water, sewers, gas, district heating, street lighting, and telecommunication, were mapped using ground penetrating radar (GPR) which was compared to historical records.  The comparison revealed major discrepancies in the historic record including thousands of meters of unknown infrastructure. For the known infrastructure large errors of geolocation were recorded.  The other ground breaking part of the pilot project was an economic analysis of the costs and benefits of applying GPR to detect the location of underground infrastructure. The analysis estimated that the return on investment is about €16 for every euro invested in improving the reliability information of underground infrastructure.

Initiatives towards national digital twins of infrastructure

Knowing where things are underground has become important enough that in several countries around the world; France, the Netherlands, Singapore, and the U.K. initiatives to improve our knowledge about the location of underground infrastructure are already underway.

In France a nation-wide multi-billion euro project was initiated to improve the quality of the location information about France's underground utility infrastructure. A 2013 presidential decree mandates that by 2019 the location of critical underground infrastructure in urban areas will be mapped to 40 cm or better. (According to the decree critical infrastructure does not include telecommunications or water and sewers.) Furthermore by 2026 the location of all of the nation's infrastructure will be mapped to 40 cm. The decree also specifies the liability for hitting underground infrastructure that shares the responsibility between the utility owner and the the contractor depending on the precision of the information available about the underground facility.

Underground Netherlands

But more importantly, as a potential model for the cost-efficient sharing of information about the underground, the Netherlands has embarked on a national program supported by legislation and standards to expand the collection and sharing of data about the subsurface. In 2015 a new law was passed by the States General in the Netherlands which created the Basisregistratie Ondergrond (BRO) or Key Registry for the Subsurface which is open and accessible to all citizens of the Netherlands. The law mandates that if you excavate or drill you have to share your data with the BRO registry. In addition if when using the data in the registry you find something is incorrect you have to report it.  The Key Registry for the Subsurface (BRO) came into force in January, 2018. The BRO registry consists of 26 data types, which will become mandatory in installments over five years and all of which include location. On 1 January 2018,it become mandatory to report the first three data types, geotechnical surveys (CPT), groundwater monitoring wells and soil drilling sample profiles. On 26 June 2018, this data became publicly available via the Dutch open data portal PDOK.  Work is underway to implement data models for the remaining data types including cables and pipes for which the Dutch standard IMKL will be used.  All 26 will become mandatory by 2022. The information model cables and pipes (IMKL) is a Dutch standard data model for the exchange of data about underground infrastructure and is founded on the INSPIRE model for cables and pipes, in which location is specified in the Open Geospatial Consortium (OGC) GML format.

As a result of the conjunction of the release of three UK government reports, Industrial Strategy Building A Britiain for the Future, Transforming Infrastructure Performance, and Data for the Public Good, the transformation in how infrastructure is built, managed and operated in the UK has made a national digital twin a key concept for the UK government. A national digital twin includes above and below ground assets. It is based on the foundation concept that a digital model is equally important as the physical assets.  Combining above and below-ground information into one national single data model/data exchange framework will allow industry to share business developments and innovation activities.  Project Iceberg is an exploratory project undertaken by the British Geological Survey, Ordnance Survey and the Future Cities Catapult to investigate ways to integrate data and services relating to the underground with other city data. To date two reports Market Research into the Current State of Play and Global Case Studies and Defining the problem space for an integrated data operating system above and below ground have been published and are publicly available.The medium term objective is to take these concepts forward with project partners to develop new digital data demonstrator projects.

Underground Singapore: Straits Times

In Singapore the Urban Redevelopment Authority is planning to have a masterplan of Singapore's underground spaces ready by 2019. It will be released as part of the next Master Plan guiding Singapore's development in the medium term.

In the U.S. an initiative has just been started to create a national infrastructure map which would include subsurface infrastructure. In June of this year a special summit was convened at Arizona State University which included leaders in public administration, infrastructure development, geography, GIS, and data integration/open data. The objective of the meeting was to review the current state of location-based information on national infrastructure including underground, examine efforts to integrate location based data systems across jurisdictions, understand the perspective of stakeholder communities, identify strategies for more systematic access to data at the national scale, and discuss the role of government to implement such strategies.

Summary

Mapping the underground is an essential component of creating a national digital twin. Recent technical development including the ability to collect data at highway speeds that can be used to create a 3D model of underground infrastructure, software that can extract relevant 3D underground information from consumer digital photography and video capture, novel ways to share information about underground infrastructure during construction at no significant additional cost to contractors, and progress on standards for sharing information about the subsurface are bringing us close to being able to cost-effectively create a digital twin of underground infrastructure at the municipal, regional and national levels increasingly feasible.

A key motivation for 3D mapping the subsurface is the risk to construction projects of unknown or poorly mapped underground infrastructure. Every year workers and the members of the public are injured or killed as a result of unforeseen underground utility hits. Underground utilities are the number one cause of delays in highway construction projects. The economic cost to the U.S. economy of not knowing where underground utilities is estimated to be in the tens of billions of dollars. A number of studies have concluded that the ROI of improving the accuracy of location of underground infrastructure is significant, with estimates of the benefits ranging as high as $21 for every dollar invested.

Singapore, the Netherlands and France have adopted policies leading to the development of accurate digital models of their national underground infrastructure.  Putting this in the context of smart cities, where data in the form of a national digital twin becomes just as important as physical assets, it is essential that a national digital twin include above and below ground assets. Fundamental to this process is rethinking value, not just the value of a finished building or infrastructure asset, but over the entire lifecycle of an infrastructure asset.  The development of a national digital twin has recently been enunciated by the UK government. In the United States an initiative has just been kicked off in March 2018 to develop a national infrastructure map including underground infrastructure.  Taken together - the public awareness of the problem, the technical advancements in underground detection in the last few years that are making cost-efficient 3D mapping of the underground increasingly feasible, and the growing recognition of the the benefits of accurate 3D maps of underground infrastructure – have led to accelerated momentum to create a digital twin of the subsurface at the municipal, regional and national levels.

Call for participation to develop model for sharing location of underground infrastructure

Public safety is at risk whenever underground excavation occurs. For example, when a gas line penetrates sewer lines during trenchless horizontal drilling, there is a risk of explosion when a sewer line subsequently backs up and an attempt is made to clear the obstruction with an augur. During construction workers accidentally rupture a high pressure gas line leading to an explosion.  It requires hours to identify the source of an explosion in a residential neighbourhood which caused fatalities.  A project to extend a light rail transit system encounters many unknown underground utilities during construction which delay the project by one and half years. The number one cause of highway construction delays in the U.S. is missing or inaccurate information about the location of underground utilities. To address the risk of liabilities associated with unknown or inaccurately located underground utilities, contractors regularly increase bid costs by a minimum of 10-30%.

All of these things can be prevented if an accurate 3D digital twin of the underground were available. To address these challenges three organizations sponsored a concept development study by the Open Geospatial Consortium (OGC) for underground information with a main outcome being an Engineering Report. The sponsors of the OGC Underground Concept Development Study were:

• Fund for the City of New York - Center for Geospatial Innovation

• Singapore Land Authority

• Ordnance Survey (UK)
  

The Engineering Report documents the progress made by the OGC and its members to build a complete situation assessment and develop a conceptual framework for action to improve underground infrastructure data interoperability. The report also identifies the most important steps to be taken next in order to develop the necessary data standards.

This led to the development of draft versions of a Model for Underground Data Definition and Interchange ("MUDDI Data Model") and an RoI (return on investment) model Cost Benefit Assessment of Subterranean Information Management. MUDDI is being developed as an interchange and integration model to support a range of critical underground infrastructure and environment information that is currently held in a wide and disparate variety of forms.

To review and assess MUDDI through actual implementations, the OGC has issued a call for participation for a MUDDI ETL-Plugfest Workshop to be held July 24-25 at the New York City Offices of the Fund for the City of New York (FCNY). An ETL-Plugfest focuses on extract, transform, and load practices and tools for the translation of data between differing formats, schemas, and environments.

National digital twin is key to the UK government's transformation of the construction industry

The conjunction of the release of three UK government reports, Industrial Strategy Building A Britiain Fir for the Future, Transforming Infrastructure Performance, and Data for the Public Good, has set the stage for a transformation in how infrastructure is built, managed and operated in the UK. National digital twin is now a key concept for the UK government and BIM and geospatial are foundational for this digital replica of the national physical infrastructure.

Mark Enzer, chief technology officer at Mott MacDonald UK and proposed chair of the Digital Framework Task Group (DFTG) which is tasked to drive these reports’ implementation forward in his keynote at Geo Business 2018 in London, explained that there will be a greater focus on existing infrastructure.  Digital abundance where the cost of everything digital has dropped dramatically over the last couple of decades has transformed many industries from banks to airlines automobile manufacturing. Construction represents about 10 % of GDP and construction productivity has plateaued over the past 40 years whereas general industrial productivity has doubled. Mark pointed out that the UK has primarily mature infrastructure to maintain and operate and construction has to change to reflect this.

Based on the concept that data in the form of a national digital twin is just as important as physical assets, digital delivery and physical delivery go hand in hand. Key to digital delivery are BIM, geospatial, a common data environment, and asset information management.  A national digital twin would include above and below ground assets.

Managing this data is about making sense of it for better decision making. Fundamental to this process is rethinking value, not just the value of a finished building or infrastructure asset, but output per ‎£ over the entire lifecycle of an infrastructure asset. This means moving beyond BIM Level 2 to full lifecycle BIM including operate and maintain.

A coordinated digital transformation landscape is required to achieve this digital transformation strategy for economic transformation. He sees the many organization in the UK infrastructure and construction sector coalescing around ICG, representing Highways England, Network Rail, Crossrail, Highspeed Rail 2 (HS2), Heathrow Airport and others and Centre for Digital Built Britain (CDBB) representing the UK BIM Alliance and others to enable this to happen.  The Centre for Digital Built Britain is a partnership between the Department of Business, Energy & Industrial Strategy and the University of Cambridge to deliver a smart digital economy for infrastructure and construction for the future and transform the UK construction industry’s approach to the way the UK plans, builds, maintains and uses its social and economic infrastructure.

First step toward developing a national infrastructure map to enable smart investments

In its 2017 Infrastructure Report Card, the American Society of Civil Engineers (ASCE) assigned an overall grade of D+ to U.S. infrastructure. Addressing the problem of bringing national infrastructure up to a standard where it becomes an enabler of economic expansion rather then a drag on the economy is hindered by a lack of systematic information readily available to decision makers on the nation’s inventory of infrastructure assets, their location, and condition. Geospatial data and technology are key to providing an accessible platform making it possible for decision makers to have a uniform view of the nation's infrastructure.  Access to this data is constrained by a variety of factors including the lack of interoperability across systems and proprietary rights. Also, data is often not comparable.  Consequently, infrastructure challenges cannot be assessed in an integrated way or communicated effectively to the relevant decision-makers and constituencies.

Initiative to create a national infrastructure map

To begin to address this challenge National Academy of Public Administration (NAPA), Arizona State University (ASU), the American Geographical Society (AGS), and the National Academy of Construction (NAC) partnered to put on an event to convene leaders in public administration, infrastructure development, geography, geospatial, and data integration/open data. Entitled "Building a Shared Map of the Nation’s Infrastructure to Enable Smart Investments" , the event was held at Arizona State University in Tempe, Arizona.

By way of background NAPA is a congressionally chartered, non-partisan organization with a mandate to make government work better.  After the last presidential election NAPA decided to get out of Washington and talk to people in different levels of government across the country.  NAPA talked to local government people in Sacramento, Austin, Syracuse, and in Virginia and came to the conclusion that there is a need for a national strategy for infrastructure.  Infrastructure is local, but data with the granularity that is required is not available.  Data is uneven, is not integrated and is basically not interoperable.  The challenge is not technical, but one of leadership and effective public administration.  The need for a national strategy was reinforced by the Trump administration's infrastructure plan which appears to only provide federal support for state initiatives.

State of U.S. infrastructure

Introduced by Terry Gerton, President and CEO of NAPA and Jonathan Koppell, Dean, College of Public Programs, Arizona State University, the keynote was delivered by Greg DiLoreto, Former President, American Society of Civil Engineers, who provided an overview of the challenge from the perspective of the latest American Society of Civil Engineers report card.

• Aviation D : Two million passengers are served daily bu U.S. airports, but congestion is growing and it is expected that 24 of the top major airports may experience "Thanksgiving-peak traffic volume" at least one day a week.
• Bridges C+ : The U.S. has 614,387 bridges of which 9.1 % or 56,000 were structurally deficient in 2016.
• Dams D : 17% of U.S. dams or 15,498 have been identified as high-hazard potential dams.  Nationwide, 73 dams failed between 2010 and 2015.
• Drinking water D : There are an estimated 240,000 water main breaks in the U.S. every year and 6 billion gallons of treated drinking water are lost every year.
• Energy D+ : There were 3,571 power outages in the U.S. in 2016. Most electric power infrastructure has exceeded its life-expectancy.
• Hazardous waste D+ : There are over 18,000 hazardous waste sites in the U.S. 53% of the American population live within three miles of a hazardous waste site.
• Inland waterways D : The U.S. has 25,000 miles of inland waterways. 49% of vessels experience delays across the waterway systems.
• Parks and recreation D+ : Seven of ten Americans use park facilities on a regular basis. About $11.9 billion in maintenance has been deferred by the National Park Service because of underfunding.
• Rail B : The U.S. rail industry invested $27.1 billion in infrastructure in 2015. Rail was one of the brighter spots on the report card.
• Roads D : America's road are often congested resulting in Americans spending 6.9 billion hours, or 42 hours per person, stuck in traffic every year.  The The Pacific Institute for Research and Evaluation (PIRE) estimates that over half of the deaths that occur on America's roadways can be blamed on poor road conditions.
• Schools D+ : There are about 100,000 schools in the U.S., serving 50 million K-12 students every day.  About a quarter of public schools are rated to be in fair or poor condition.
• Solid waste C+ : Americans generate 258 million tons of municipal solid waste every year of which 53% ends up in landfill.
• Transit D- : Transit in America is losing ridership in some regions, the result of chronic underfunding which has generated a $90 billion maintenance backlog.

The state of much U.S. infrastructure is poor because of chronic underfunding.  Greg pointed out that China spends more on its infrastructure than North America and Europe combined.  The only real bright spot in the 2017 report card is rail.  There are slight improvements in seven of the categories the ASCE tracks.  But in three sectors the ASCE's lower grade reflects a worsening assessment. Transit, solid waste and parks and recreation have seen a decline since the last report in 2013.  For 2016-2025 the ASCE estimates that there is an investment gap of about $ 2 trillion.  This is the difference between the actual allocated funding and what is needed. 

 

Substandard infrastructure reduces the income of every American.  The ASCE estimates that failure to act will result in a loss of $3.9 trillion to the U.S. economy, which translates to a loss of $3400 per year per family.  The ASCE estimates that to bring U.S. infrastructure up to an acceptable standard would require investment equivalent to a latte per day, or $4 per day, per family. 

 

To fund infrastructure properly requires increasing investment from the current 2.5% of GDP to 3.5%.  Greg emphasized that this will require increased investment from all levels of government and the private sector.  Greg suggested some ways to improve how infrastructure is funded. First, congress should raise the federal motor fuel tax.  This would increase funding of federal highways through the Highway Trust Fund.  Secondly,  the contribution from the federal government should be increased.  When the interstate highway system was built during Dwight D Eisenhower's administration, the federal government provided 90% of the funding.  Since then the funding ratio has been inverted.  Greg also emphasized that when highways are built, full lifecycle cost analysis needs to become standard practice.

 

The good news is the message about the importance of infrastructure to the economy is getting out.  A recent survey found that 87% of Americans think that the federal government should increase its investment in infrastructure.

Deep learning enables automated extraction of building footprints and road networks from satellite imagery

Automated feature extraction from satellite imagery has made major progress in the last year. Accurate building footprints extracted from high resolution satellite imagery are available from companies such as Ecopia (which has just announced a partnership with DigitalGlobe).  Also NVIDIA has demonstrated the ability to automate detection of many road networks using sophisticated algorithms and multi-spectral high resolution imagery.

Building footprints

SpaceNet is a repository commercial satellite imagery and labeled training data being made available through Amazon Web Services at no cost to the public to foster innovation in the development of computer vision algorithms to automatically extract information from remote sensing data.  It is the result of a collaboration between DigitalGlobe, CosmiQ Works, and NVIDIA.  Two competitions for automated building footprint extraction from high resolution satellite imagery and a third for road network extraction have been held based on SpaceNet data. 

The first SpaceNet Challenge was launched in November 2016.  Imagery of Rio de Janeiro from the WorldView 2 satellite at 50cm resolution using eight spectral bands was made publicly available on Amazon Web Services (AWS).  42 developers competed to create algorithms that extract building footprints from the imagery.  The source code of the winning algorithms was made available on the SpaceNetChallenge GitHub repository. 

The winning implementation was developed by Brazilian developer using an implementation based on random forests with brute force polygon search and did not leverage deep learning frameworks.  The organizers' conclusion based on the results of Round 1 was that automated building footprint extraction remained a challenge.

Round 2 winner. Most building footprints are precisely detected. Number on each building footprint represents the overlap between the prediction and the ground truth.

Round 2 of the SpaceNet challenge involved higher resolution 30 cm imagery from WorldView 3 for four cities across the globe. The competition challenged developers to improve performance from the first competition using the higher-resolution imagery and more geographically diverse training data samples. SpaceNet on AWS held imagery and building footprints for Las Vegas, Paris, Shanghai, and Khartoum with over 180,000 buildings181,619 footprints.  In addition additional bands were made available (panchromatic, 3-band pan-sharpened, and 8-band pan-sharpened).

Results from Round 2 winning solution: top left: Las Vegas, top right: Las Vegas, lower right: Khartoum, and lower left: Paris. Blue outlines: ground truth, green outlines: correct identifications, red outlines: false positives and yellow outlines: false negatives.

The winner of Round 2 applied a  deep neural network model developed originally for medical image segmentation called U-Net.  Successful training of deep networks requires many thousand annotated training samples.  Training was aided significantly by augmenting the imagery with OpenStreetMap data.  The winning solution used OpenStreetMap layers and Worldview-2 multispectral layers as the input of the deep neural network algorithm.  In Las Vegas, the winning solution was outstanding in outlining suburban homes, but also did well in identifying odd-shaped buildings. In Khartoum, the solution was very good in identifying the footprints of stand-alone apartment complexes, but had trouble with small buildings, with building footprints that are close together, and with horizontally large buildings.  One of the problems was inconsistent ground truth which affected training. The solutions offered by the developers in the competition were sufficiently successful that the organizers concluded that the winning algorithms achieved performance with potential for automated mapping tasks such as keeping maps up-to-date and to assist first responders during natural disasters.  The source code for the winning implementations can be found in the SpaceNetChallenge GitHub repository.

DigitalGlobe has just announced a partnership with Ecopia, which has established an automated process to create building footprints quickly and at scale by leveraging DigitalGlobe's Geospatial Big Data platform (GBDX) and advanced machine learning in combination with DigitalGlobe’s imagery library. The two companies plan to automatically extract accurate 2D building footprints globally, then refresh the datasets periodically to find and track changes over time.  These datasets would be valuable to municipal governments for permitting purposes and for first responders after disasters.   Ecopia is developing a database for all of Australia (7.6 million square kilometers) which will include every building with a roof area greater than 9 square meters.  The database will contain building footprints for about 15 million buildings. The database includes linkages to other datasets including geocoded address, property data and administrative boundaries. Ecopia intend to update the dataset regularly to be able to track changes, especially urban sprawl.

Road networks

In November 2017 CosmiQ Works, Radiant Solutions, and NVIDIA announced a third round competition, “Road Detection and Routing Challenge” to explore automated methods for extracting map-ready road networks and routing information from high-resolution satellite imagery.  

Detecting road networks in Paris. Photo image, ground truth, and the results achieved with three deep learning algorithms

NVIDIA has released the results of several deep learning algorithms that illustrate just how difficult identifying road networks is and the sophistication of the available tools.  These include using multi-band spectral imagery to identify the material properties of the road surface itself (asphalt, gravel, packed earth) and using more sophisticated processing including conditional random fields (CRF), percolation theory, and reinforcement learning. 

Detecting road networks in Las Vegas. Photo image, ground truth, and the results achieved with three deep learning algorithms

NVIDIA utilized high performance GPU compute resources provided by the NVIDIA GPU Cloud and Amazon Web Services.