Construction projects are extremely complex and interdependent endeavours, and the potential for inefficiency and risk – which eventually lead to high project costs and delays – grows exponentially with project scale. Here, we will focus on the applications of the SKY ENGINE AI platform to a development of AI models for workers safety monitoring, on-site inspection, audit, and infrastructure management in the construction sector to reduce operating costs and optimize the efficiency of these jobs. We will explain the analytics and inference capabilities of the SKY ENGINE AI platform's Computer Vision (CV) solutions for heavy industries, and the capacity to assist construction companies in determining which combination of factors is most likely to result in an incident in the future.
Efficiently managing complexity on a construction site with safety assurance
Numerous construction companies have increased their AI-driven automation initiatives in response to rising risk, stressed supply chains, and shrinking profits. Typically, businesses have concentrated on establishing operational efficiencies via the use of technology to streamline processes and procedures, but the data available on site is frequently used in a suboptimal way.
What if you could highly increase workers safety on a construction site and boost your chances of completing a project on time and under budget by leveraging synthetic data simulated and generated in SKY ENGINE AI platform for AI models training? Obviously, the number of on-site data that you previously just saved for future reference will also serve further AI models enhancement. Such solution build in the SKY ENGINE AI platform can assist engineering and construction (EC) companies in optimizing their decision-making and driving project success by proactively uncovering new insights from construction site data.
Consider the following scenario: it is 1 p.m. on a hot July day, and team members are working at height. To ensure worker safety, vital signs of team members are checked. A worker grows disoriented as his body temperature rises, putting him in risk of falling. In this case, such smart AI solution notifies the worker and site team of the hazard so that they may take appropriate action.
What about utilizing a drone to hover above team members to assist with high-risk or difficult work? The Adaptive Construction Intelligence solutions can be created in the SKY ENGINE AI platform to be ready for use with drones. The drone might be placed into position shortly before work begins to check settings, prompt the worker with safety questions, and provide instructions through an e.g. earpod. Given the worker's experience and track record, the drone would monitor the worker's status as well as the site circumstances, predicting potential complications. If the hazards grow, the drone may alert the worker or sound an alarm. These types of solutions can also be seamlessly integrated with an existing network of surveillance cameras installed on the construction site to further minimize the overall inspection system's cost.
Injuries and deaths on the job are much too prevalent in the construction industry. While precise figures vary by region, it is believed that tens of thousands of construction workers are still wounded on the job each year. Falls, electric shocks, being struck by an object, or getting stuck between objects are the top safety concerns, which have remained virtually constant for some years. Despite decades of industry effort and regulatory rigor, building remains one of the world's most hazardous vocations.
This begs the question, why has building site safety improved so little in recent decades? The answer is... complexity. Construction endeavours are very complex systems of interactions, interdependencies, and relationships. Some consider the construction business to be the most intricate of all industries.
Three major categories very important from safety and project management standpoint can be effectively tracked and these are Manpower, Machines and Materials.
Figure 1 – SKY ENGINE AI platform enables development of several computer vision AI models and applications for people tracking, 3D pose estimation, machines utilization tracking, materials flow and more to gain insights into complex construction site environment, track progress, ensure safety and predict risks before they will become severe.
The AI models developed in the SKY ENGINE AI platform can cover wide range of scenarios and job-sites from infrastructure, building works to property management. Effectively that's the true enabler of continuous monitoring and active learning system. SKY ENGINE AI provides construction specified AI algorithms from ground research to cover several scenarios that may occur on the construction site every day. Our proprietary algorithms can work under extreme environment in different types of job site. We understand customer operation scenarios and the AI follows as well. Personnel can be tracked to ensure their safety in various situations. We can create a digital twin of construction site where we may track the changes and progress. Equipment can be monitored for productivity and its utilization and overall optimization of the on-site operations. All such data can serve to analyse trends, look for correlations, and identify potential problems.
Figure 2 – AI-driven analysis of a construction site including vehicles and machines tracking to assess their utilization; workers monitoring for safety assurance and PPE compliance. Despite the complexity of such scene the SKY ENGINE AI platform can effectively drive the accurate AI models production to provide support to the construction companies offering insights not available with conventional monitoring.
Smart Construction Sites in action – turning data into insights
Let’s take closer look on this construction site scene above. It’s obviously complex with several machines and workers involved. That site evolves very dynamically and manually capturing all the potential issues might be very difficult and time consuming. However, the AI models can effectively monitor entire area including the 3D motion of the machines (position and distances), compliance to the PPE rules, workers proximity to the machines – you can spot blue boxes over the workers with no issues and the red boxes for the others that are non-compliant with PPE measures or those being too close to the moving machines which poses obvious hazard to them. This unlocks proper identifcation of any potential risks before they escalate to injuries and health-related issues or even fatalities. Stopped vehicles are marked with cyan color and the vehicles in motion are green. Monitoring of working machines enables correct assessment of their level of utilization which can serve better redistribution of the equipment even next day. The AI algorithms can also be trained to count machines and workers visible in the scene, and to identify high-risk zones on the job site. If a worker approaches one of those zones, the system will alert the project manager to move the worker away from a danger.
Everyone in the construction industry is responsible for safety, including employees, managers, and leaders, as well as regulators. Our human brains, on the other hand, have a limited capacity to manage and interpret complexity, and we struggle to process enormous volumes of information at the same time. We also bring a range of cognitive biases to work, such as underestimating dangers or overestimating the efficacy of our own mitigation techniques. You may believe that relying on highly experienced safety managers is the solution. Even the most experienced safety managers, however, learn from just a hundred or so events over the course of their careers. Wisdom and good intentions are insufficient to move the sector forward. In fact, since there are so many levels of complexity involved, it is nearly impossible for humans to recognize all potential threats and conduct appropriate risk assessments.
This is where SKY ENGINE AI helps in making construction workplaces safer. The SKY ENGINE AI’s core strength is generating and processing vast amounts of data quickly and finding meaningful patterns. Imagine using information such as site conditions, worker experience, and project size to help predict the likelihood and severity of future safety issues.
Figure 3 – Construction site safety assurance. (Left) Workers at height – PPE check: hard hats, harnesses, high-vis vests, safety ropes. (Right) PPE compliance tracking (workers), equipment utilization and counting and authorized access monitoring including licence plates (vehicles, excavators).
Personnel safety monitoring
Additional demos of a personnel safety computer vision solutions based on the simulated synthetic data for the AI models training are shown above. These demos showcase tracking the workers operating at height and their compliance with safety PPE rules. The health and safety of everyone on a project is always top priority. The emerging signs that can alert organizations of potential safety issues are all too obvious after an accident occurs.Capturing and analyzing data in the SKY ENGINE AI platform enables predictive intelligence that can help to detect potential issues by streamlined analysis of the images and video feeds from the jobsite, data from sensors. Moreover, safety reports, correspondence, training logs, and past incidents can enhance the predictive analytics and serve building comprehensive and accurate alerting system. This can provide organizations with an early warning of potential emerging safety issues so that they can address them proactively. In addition, it can reduce expenditure related to any potential insurance claims. If a recurring safety violation occurs, or workers enter danger zones frequently, this data is collected and can be assessed at a later date. This provides managers with the information needed to remedy any issues. Perhaps more safety training is needed, or perhaps better signage is required around a specific location on the construction site. These safety metrics can be monitored and measured for an accurate understanding of safety hazards on site and per each worker.
Predict project risks
When suppliers fail to deliver materials on schedule, projects fall behind. When data streams and statistics indicate that delivery vehicles are running late, you may take corrective action with suppliers. In order to stay on track it is critical to identify possible problems before they materialize. Real-time data identifies bottlenecks, allowing you to react quickly and avoid future delays. It also brings the ability to conduct remote inspections, which can reduce delays at various phases of construction.
If construction progress comes to a halt abruptly, you'll want to know why. When data streams and AI-driven data analysis identify situations when machines or personnel are not performing as expected, you may call the subcontractor and get the production back on track, avoiding cost overruns. For project cost control, the SKY ENGINE AI models can track material arrival and departure insights.
Figure 4 – On-site operations intelligence and suppliers tracking using AI models created with procedurally simulated/generated synthetic data of environment, equipment, materials, etc. in the SKY ENGINE AI platform with insights data for predictive analytics and continuous progress awareness.
The AI solutions for construction site analytics can be integrated with Control/Command Center to enable Safety Managers getting live insights from several construction zones of interest at once. These insights can be provided in two modes (see Figure below): LIVE mode that is typical display of all the available video streams from all the selected cameras. And ALERT mode – that shows only events that need further investigation (tiny fraction of total stream), decreasing operator's fatigue and event misses. This can be easily integrated with any video management systems.
The AI solutions created in the SKY ENGINE AI platform can be employed to help Construction Enterprises:
- Understand and analyse workers safety & productivity,
- Enhance operational workflow,
- Provide insights in case of insurance claims,
- Perform statistical analysis and identify construction sites and/or teams where safety rules are frequently breached and,
- Help ensuring environmental compliance.
Several video analytics modules can be created in the SKY ENGINE AI platform for comprehensive monitoring of the construction site (see below) for personnel safety, operations, productivity and environmental mapping.
Adaptive Construction Intelligence®
These solutions may be used to assist in identifying event risks, such as their likelihood, immediate cause, severity, and potential injuries. Incidents can be foreseen, managed, and, at best, prevented entirely with the proper amount of readiness. Managers may be advised of the most likely events, given the site circumstances and planned work packages, with a high degree of accuracy. This assists site leaders in discussing concerns at pre-job workplace safety meetings, reinforcing specific safety measures, and recommending additional training and supervision assistance. Work can also be transferred to better-suited personnel and, if required, work packages can be restructured to reduce hazards. In addition, a real-time data can be used to enable continuous learning on a construction site. For example, it might detect employees who are not wearing hard hats or safety gear, as well as the absence of safety rail guards or fall-protection equipment. Any changing conditions (team- and site-specific) can be an input into the algorithms indefinitely for analysis and action.
Let us know about your cases and get access to the SKY ENGINE AI platform or get a tailored AI models or synthetic datasets for your construction site computer vision applications. As we support much more industries a broad range of data and AI models customization is available even for specific sensors and environments.