Computer Vision in Construction

Part
01
of six
Part
01

Computer Vision Service Providers Part 2

Five of the top service providers of computer vision for the construction industry include Indus.ai, OpenSpace, Astralink, Avvir, and SenSat. Indus.ai, Astralink, and Avvir are using computer vision technology to monitor construction sites and track building progress. OpenSpace and SenSat have developed computer vision-powered systems to capture and map construction sites.

Indus.ai

OpenSpace

Astralink

  • Astralink is an augmented reality platform powered by computer vision that helps construction companies prevent errors in real-time by comparing the actual state of the construction sites with the design plans.
  • AEC Business has named Astralink as one of the startups that are in the “vanguard of turning AI into a value-adding technology for construction sites.”
  • Astralink has raised $1 million in funding to date.
  • The company is based in Tel Aviv, Israel.
  • It was founded in 2015.

Avvir

SenSat

Research Strategy

An extensive examination of lists that featured computer vision companies in the construction industry did not provide any companies that have appeared in at least two lists. The lists published by AEC Business, Built In, AI Startups, and others have provided only three additional companies (Indus.ai, OpenSpace, and Astralink) that appeared on one list. All three companies are venture-backed startups. Thus, we have selected two more companies (Avvir and SenSat) that have received media coverage recently and raised venture capital funding.
Part
02
of six
Part
02

Computer Vision Service Providers Part 1

Five of the top service providers of computer vision for the construction industry include Reconstruct, OnSiteIQ, INTSITE, Disperse, and Doxel. Reconstruct, OnSiteIQ, and Disperse offer computer vision-powered construction management software. INTSITE and Doxel provide computer vision-driven machinery and autonomous devices respectively.

Reconstruct

OnSiteIQ

INTSITE

Disperse

Doxel

  • Doxel is a computer vision and artificial intelligence-powered system that uses autonomous devices to monitor commercial construction projects, inspect the quality of the work installed, and measure the installed works in real-time.
  • The Robot Report has named Doxel as one of the companies using construction robotics that are disrupting the industry.
  • Doxel has raised $4.5 million in funding to date.
  • The company is based in Palo Alto, California.
  • It was founded in 2016.

Research Strategy

An extensive examination of lists that featured computer vision companies in the construction industry did not provide any companies that have appeared in at least two lists. The lists published by Built In, AEC Business, AI Startups, and others have provided only three companies (Reconstruct, OnSiteIQ, and INTSITE) that appeared on one list. All three companies are venture-backed startups. Thus, we have selected two more companies (Disperse and Doxel) that have received media coverage recently and raised venture capital funding.
Part
03
of six
Part
03

Computer Vision Competitive Landscape: Construction

Based on the amount of funding raised, the top three service providers of computer vision for the construction industry are OpenSpace ($17.5 million), Disperse ($16.7 million), and Sensat ($15 million). OpenSpace's Vision Engine platform and Sensat’s Mapp are mapping systems powered by computer vision. Disperse’s computer vision-driven platform helps construction companies track building progress.

OpenSpace

  • OpenSpace has raised $17.5 million in funding to date. The company’s last funding round was a $14 million Series A round in August 2019.
  • The technology developed by OpenSpace is based on the PhD work of its co-founders (Jeevan Kalanithi, Philip DeCamp, and Michael Fleischman) at the Massachusetts Institute of Technology.
  • The main target market for OpenSpace's Vision Engine platform is the construction industry.
  • The foundational technologies behind the automated construction site image capturing and mapping system include computer vision, 3D reconstruction, and big data visualization.
  • OpenSpace's clients include The Beck Group, DPR Construction, Hathaway Dinwiddie, and others.

Disperse

  • Disperse has raised $16.7 million in funding to date. The company’s last funding round was a $15 million Series A round in October 2019.
  • Chief Executive Officer Felix was “raised on construction sites from a young age” and co-founded Disperse while pursuing a master's in digital innovation at the London School of Economics.
  • The main target market for Disperse’s solution is the construction industry.
  • The computer vision-powered platform helps construction companies track progress and includes features such as 360º progress feed, objective progress reports, early warning alerts, management dashboards, update scheduling, and quality and output comparison.
  • Disperse’s clients include Canary Wharf Construction, Bouygues UK, Meyer Bergman, and others.

Sensat

  • Sensat has raised $15 million in funding to date. The company’s last funding round was a $10 million Series A round in October 2019.
  • The company’s co-founders Harry Atkinson and James Dean had originally built a “prototype platform that transformed large Structure from Motion (SfM) datasets into representations of the physical world as binary numbers.”
  • The main target market for Sensat’s Mapp is the construction industry.
  • Mapp is a computer vision-powered simulated reality program and supports construction projects at all stages, which include planning & pre-construction, build & project management, and manage, maintain & operate stages.
  • Sensat’s clients include Highways England, Mace, Murphy, and others.
Part
04
of six
Part
04

Computer Vision Insights: Construction

Computer vision is being used in regions with the fastest growth in real estate development, especially in China and India, and in the near future it will become more widespread in the Middle East, especially in the UAE. Insights surrounding the use of computer vision in construction include for site safety, resource monitoring, infrastructure assessment, and measuring construction progress.

Computer Vision and Global Regions With the Fastest Growth in Real Estate Development

  • According to PwC, in 2020 a lot of real estate development will be taking place in Asia, Africa, the Middle East, and Latin America. While the Asian market will be the fastest growing market in 2020, sub-Saharan Africa will be the second fastest growing market. The report also highlights specific countries a key players in this development, such as China, Abu Dhabi, India and South Korea. The report also notes some developed nations among those that will see high rates of real estate development due to increasing population size, such as the U.K. (London), and the U.S. (New York City).
  • Overall, the APAC region has a lot of growth potential when it comes to computer vision, especially in China, India, Japan and South Korea. As of 2017, China held one-third of the global computer vision market, even surpassing the market share of the United States. The Chinese government expects China to become the global leader in the overall A.I. market by 2030. There is a lot of computer vision development taking place in India, as 50% of AI startups in the country are working in the areas of computer vision-based AI technologies.
  • There is increasing adoption of A.I. technologies in the Middle East and especially in the UAE, due to large part of the smart city initatives of major UAE cities, such as Dubai. Spending on AI technologies in the MEA region reached $290 million in 2019, with a growth rate of 43%. As such, many AI vendors are opening offices in the UAE, and many new vendors are being founded here as well, including computer vision company, Hazen. However, a 2019 report published by Microsoft notes that: "While companies acknowledge the significant potential of broader, more advanced AI technologies such as computer vision [...] they are currently not in common use by companies in the MEA region." Only around 16% of companies surveyed in the region found computer vision to be useful in their company's AI arsenal (the lowest ranked AI technology of the survey). Despite this, ongoing efforts in this region will prime the region to readily deploy these types of advanced AI solutions in the future.
  • Alongside China, the United States is investing the most in AI technologies globally, with EU countries gaining a lot of ground. These technologies include computer vision.

Computer Vision Usage in Construction

#1: Improving Construction Site Safety Using Computer Vision

  • According to a 2018 academic report, computer vision has been introduced into the construction industry in recent years, with a primary use case to improve occupational health and safety.
  • The unsafe behaviors of construction workers is a key contributor to 80% of accidents. Safety in the construction industry can be improved via the identification, assessment, and modification of these behaviors.
  • In 2017, Yale developed a framework using computer vision that can detect specific types of unsafe actions and behaviors occurring on construction sites, while other researchers developed similar frameworks, such as Fang's 2018 computer vision algorithm that can detect construction workers who aren't wearing harnesses.
  • Despite these innovations, there are a number of theoretical and technical challenges preventing their widespread implementation at the current moment. However, these advancements are noted as positive steps in terms of laying a "fundamental basis for more powerful health and safety assessment," as further research is being carried out.
  • Despite these challenges, computer vision technologies are already being used to improve the safety of construction sites by doing things like identifying whether workers are wearing hardhats, safety jacks/gloves/footwear and other types of personal protective equipment.

#2: Use of Computer Vision Technology in Infrastructure Assessment

  • Computer vision technologies can operate in conjunction with remote cameras and drones (UAVs) to "offer promising non-contact solutions to civil infrastructure condition assessment," according to a 2018 academic report.
  • Currently a lot of research has been being conducted in realm of development and adaptation of computer-vision technologies for inspection and monitoring in the field of civil engineering.
  • Civil infrastructure includes dams, highways, buildings, bridges, and lifeline systems. Much of today's existing civil infrastructure is already multiple decades old and "past its design life." According to the American Society of Civil Engineers, there are currently more than 56,000 "structurally deficient bridges requiring a massive $123 billion USD for rehabilitation"
  • Traditional techniques relied on for assessing civil infrastructure conditions involve visual inspections by humans judging against a set of inspection standards. However, this process is very time-consuming, involves a lot of labor, is costly, and also dangerous.
  • While some real-time inspection technologies are currently available, they also have their own set of pain points such as spatial resolution limitations, installation and maintenance issues, and low-feasibility due to costs.
  • Due to these issues and challenges when it comes to civil infrastructure assessment, "computer vision techniques have been recognized by the civil engineering field as a key component of improved inspection and monitoring," as it serves as a solution to these pain points while also providing the best of both worlds (combing the benefits of both human and technological inspection).

#3: Measuring Construction Site Progress Using Computer Vision

  • Computer vision is capable of providing feedback about construction site progress in a way that is both accurate and timely.
  • Traditionally, measuring construction site progress can be rather subjective. Additionally, many errors can occur in this process which yield less than accurate results, and the information gathered can quickly become out-of-date.
  • However, being able to accurately track task completion on construction sites is of significant importance when it comes to controlling costs, scheduling workers, and handling money flows.
  • Although advanced technologies that can play a role in measuring construction site progress are preexisting (e.g. photogrammetry techniques, reconstruction tools, and 3D modeling), these technologies "often involve intensive user interaction and have slow turnaround." On the contrary, a fully automated system based around computer vision can serve as a "valuable aid for project management, enabling tighter control and greater efficiency."

#4: Computer Vision Can Be Used to Monitor Construction Site Resources

  • A study conducted by DiXite in January 2019 aimed to identify the real-world needs of construction sites and how machine vision could be implemented to address these needs. This study found that that one implementation for this type of technology on construction sites is for the tracking of construction site resources.
  • Computer vision systems that incorporate video cameras and drones are able to track and monitor construction site activities involving resources. These activities include things like vehicle and resource movement, as well as deliveries.
  • Tracking resources in construction is important because "construction sites suffer from under-utilization of resources, lack of insights on activity and real time intelligence, [and] more data allows managers to respond quickly to incidents or optimize construction activity to increase efficiency and speed."
Part
05
of six
Part
05

Computer Vision Market Size: Construction

There is not enough information in the public domain to reliably triangulate the size of the computer vision market in the global construction industry. The construction industry does not figure among the top industries for computer vision. Available statistics only suggest that the global market size is either below or in the region of $500 million.

Helpful Findings

  • Verified Market Research projects that the size of the global computer vision market will grow from $11.88 billion in 2018 to $21.24 billion by 2026. It segments this market into the following applications: consumer electronics, food and beverage, automotive, medical, robotics, sports and entertainment, and others. This information indicates that the construction industry is not among the top markets for computer vision and is included in the 'others' category.
  • Verified Market Research estimates that the 'others' category will grow from around $400 million in 2017 to around $600 million by 2025. Since it is very likely that the construction industry is not the only industry in the 'others' category, the size of the computer vision market in the construction industry in 2017 must be below $400 million.
  • Tractica projects that the size of the global market for computer vision hardware, software, and services will grow from $1.1 billion in 2016 to $26.2 billion by 2025. It segments this market into the following applications: automotive, machine vision and robotics, surveillance and security, medical imaging, sports and entertainment, consumer applications, agriculture, and retail. This information indicates that the construction industry is not among the top markets for computer vision.
  • Markets and Markets projects that the size of the global computer vision market will grow from $10.9 billion in 2019 to $17.4 billion by 2024. It segments this market into industrial verticals and non-industrial verticals, none of which includes the construction industry. Again, the way Markets and Markets has segmented the global computer vision market indicates that the construction industry is relatively small and is not among the top markets for computer vision.
  • In a chart published by Statista, the global computer vision market is segmented into the following six application categories only: consumer, automotive, robotics and machine vision, security and surveillance, medical, and sports and entertainment. This segmentation indicates that the construction industry is not among the top markets for computer vision.
  • Grand View Research estimates that the size of the global machine vision market in 2017 was $9.95 billion. It segments this market into the following end-use industries: automotive, pharmaceuticals and chemicals, electronics and semiconductor, pulp and paper, printing and labeling, food and beverage, glass and metal, postal and logistics, and others. This segmentation indicates that the construction industry is not among the top markets for computer vision.
  • In 2018, the size of the global artificial intelligence market was around $9.842 billion, while the size of the global artificial intelligence market in construction was around $434.038 million. These details suggest that, in 2018, the construction industry accounted for around 4.41% of the global artificial intelligence market.
  • Computer vision is a branch of artificial intelligence. If it is assumed that construction's share of the global computer vision market is about the same as construction's share of the global artificial intelligence market, it follows that the size of the construction segment is around $523.908 million or 4.41% of $11.88 billion.
  • The construction industry has started exploring computer vision to improve occupational health and safety (OHS), quality, and productivity.

Research Strategy

Since the industry reports that market research firms publish typically contain market segmentation details, we started our research with an examination of available computer vision market reports. We found that several firms, including Tractica, Markets and Markets, and Verified Market Research, offer reports on the computer vision market, so we scoured these firms' reports first. These reports, unfortunately, do not provide the size of the computer vision market for the construction industry. We only learned from these reports that the construction industry is not among the top users of computer vision, and that the construction industry must be part of the 'others' category.

This initial strategy was followed by a press search where we looked for articles covering computer vision in the construction industry. This press search led us to a few related articles, including those published by DZone, Indus, and ResearchGate; however, the information contained in these articles was mostly qualitative. There was no quantitative data that we could use to triangulate the portion of the computer vision market accounted for by the construction industry.

The size of the global computer vision market is readily available, so we proceeded to look for construction's share of the market. We figured that, if we knew both the size of the global computer vision market and construction's market share, we would be able to compute the size of the computer vision market in the construction industry. We were unable to find construction's share of the computer vision market, however. All we found was the percentage of the global artificial intelligence market accounted for by the construction industry.

Global artificial intelligence market, 2018: $4.065 billion x 1.556 x 1.556 = $9.842 billion or $9,842 million
Construction's share of the global artificial intelligence market, 2018: ($434.038 million/$9,842 million) x 100% = 4.41%

There is a report segmenting the global machine vision market by industry, but similar to the global computer vision market reports, this report does not list construction as one of the segments. Additionally, there appears to be a difference between computer vision and machine vision.

All in all, there is not enough information in the public domain to triangulate the size of the computer vision market in the construction industry. We can only deduce that the construction industry accounts for only a very small portion of the computer vision market, and that this is true regardless of geographic scope (i.e., global, United States, or other regions).
Part
06
of six
Part
06

Computer Vision Trends: Construction

Four trends regarding the use of computer vision in the construction industry include utilizing computer vision to ensure safety, improve security, maintain quality and increase productivity.

Safety

  • Computer vision can help maintain safety at the construction site by monitoring workers to ensure they are wearing safety equipment like hardhats and masks, and can alert managers to workers entering dangerous areas.
  • The monitoring can be used to alert managers in real time so they can rectify the situation. Additionally, the footage can be reviewed in the case of an incident.
  • Driving the trend for this software is the high risk environment of construction sites and the high number and cost of construction accidents.
  • One such company developing and testing software for this purpose is AGMIS.
  • A Boston, Massachusetts company, Suffolk, has created a computer vision system that utilizes AI to flag potentially dangerous situations on construction sites.

Security (S2, S1)

  • Computer vision can detect unauthorized entry into work zones by identifying vehicle license plates and utilizing facial recognition.
  • Computer vision systems can also detect when anyone enters the area after closing time or behaves suspiciously.
  • Computer vision is fulfilling a need in the construction industry, as there are high instances of vandalism and theft that cost billions.
  • Deepomatic is a computer vision company that has solutions for monitoring theft and unauthorized entry at construction sites.

Maintaining Quality

  • Computer vision can be used to detect defects in construction in real time by comparing the images to work plans or outside specifications.
  • Defects can cost the construction company not only because the work will have to be fixed, but also because they can lead to huge delays and site closures. By identifying these issues in real time as soon as possible, the company can more efficiently address mistakes.
  • The technology can also be used to identify at what stage of the project an issue occurred, in order to identify what party is responsible for the issue.
  • Disperse is a new computer vision company making software capable of detecting issues at construction sites.

Increasing Productivity

  • By tracking inventory and optimizing worker flow, companies can decrease production costs and avoid delays.
  • Driving this trend is the high cost of idle time at construction sites, the need to complete projects by deadline, and the potential cost of delays due to poorly managed inventory.
  • One such company promoting computer vision solutions to increase productivity at construction sites is Indus.AI Construction Intelligence.

Research Strategy

Trends were determined based on the mention of the trend in multiple credible reports.
Sources
Sources