What are point clouds?
Point clouds are a digital 3D representation of physical objects or spaces. A point cloud comprises millions or billions of individual measurement points, each with x, y, and z coordinates.
Methods can vary to capture the cloud and the types of sensors used. Still, each point can include RGB color data or intensity information, which reflects the measurement of the laser pulse that generated the point.
How do we get point clouds?
3D scanning, which generates point cloud data, can be traced back to the 1960s. These laser scanners were developed for space and defense applications; in the late 80s, the technology was also recognized as valuable for industrial use. Since then, the technology has seen significant hardware and software improvements, and we now have easy-to-use 3D scanners that just about anyone can successfully deploy in the field.
There are two primary means of capturing point clouds: laser scanners and photogrammetry.
Laser scanners for point clouds
A laser scanner includes several different sensors and technologies. One important type is a LiDAR sensor (light detection and ranging), which uses rapid laser pulses to gather hundreds of thousands of accurate measurements per second. Most laser scanners also use an RGB camera to add color to the point cloud and an inertial measurement unit, the professional-grade relative of the accelerometer in your mobile phone.
There are various types of laser scanners, each designed for a specific range of applications. For example, you would utilize a terrestrial laser scanner (TLS) to produce a point cloud with the highest accuracy for a particular application like analyzing floor flatness, measuring beam deflection, or capturing a single object like a machine or vehicle with incredibly high precision.
Additionally, a mobile laser scanner can enable you to capture point clouds as you walk and jobsite and provide up to 4mm accuracy. This performance and speed make them ideal for most building documentation projects, reducing the overall square meter cost of capturing data. This speed also makes them suitable for mapping active job sites with accelerated schedules offering a small time window for capturing reality.
More market options include using laser scanners for various specialized applications, like capturing roadways, railways, or objects and generating wide-area topographic maps. Moreover, utilizing multiple devices together can be used to integrate numerous point clouds into one final point cloud to meet the project team's needs.
In general, laser scanners are more accurate than photogrammetry for generating point clouds.
Photogrammetry for point clouds
Less a specific tool and more of a methodology, you can produce a point cloud using photogrammetry; you would use cameras to capture a structure from all angles and then process those images with specialized software to reconstruct the 3D model of the area or structure.
Projects that leverage drones to capture point cloud data are most likely to use photogrammetry since laser scanners are typically too heavy for drones to carry.
How are point clouds used in design and construction?
Point clouds generate substantial data sets that provide a precise, accurate, and comprehensive digital picture of a real-world space, structure, surface, or object. The result offers massive value for construction projects.
Below are some typical applications for leveraging point clouds in the built world.
Progress and productivity tracking
Automating SOV reporting
Avvir's Progress software utilizes point cloud data or 360° images and video walkthroughs to analyze and automate reporting of up-to-date conditions and status of a construction project. The majority of the 360° cameras on the market will be sufficient to capture this data. A scanner like the NavVis VLX captures 3D point cloud data at a fraction of the time compared to a terrestrial scanner, and the data can be used in the Avvir platform to track the job's progress.
Once the reality capture data is uploaded, the Avvir algorithm statuses every single BIM element as Built or Not-Built based on a sophisticated comparison of the point cloud data or 360° Photos to the BIM models. Taking this to the next level, Avvir Progress can ingest a construction project schedule to automate reporting for what's been installed on the job site vs. what was scheduled to be installed on a trade-by-trade and area-by-area basis. This allows the construction project team to catch areas or trades behind schedule.
Avvir Progress analysis can also look forward and help you anticipate upcoming critical path milestones. Avvir Progress 4D data and analysis is flexibly packaged and viewed in several ways – whether in an S-Curve 4D dashboard, within PowerBI, or directly onto your project schedule to view the data in easily digestible Gantt chart format or a custom report.
The construction world is significantly jumping into 5D applications and pay apps. Reality capture can help in this arena as well. Avvir Progress 5D tracks installed value against scheduled progress for each line item in your work breakdown structure (vs. schedule of value). In addition, Avvir Progress 5D helps validate and expedite subcontractor billings/pay apps with reality capture data.
Improving quality assurance and control
Understanding the accuracy of installed items (vs. how they were planned in the BIM) is critical to keeping construction mistakes and project costs down. Today there are typically two ways to verify if work is installed correctly on the job site: manual vs automated.
The manual verification method encompasses importing the point cloud and BIM model into a platform like Navisworks and manually comparing the two for any discrepancies. The challenge with this method is that this process is incredibly tedious and time-consuming, and the manual effort leaves significant room for error. When a deviation is identified, it is often too late to take meaningful action.
The Avvir team identified this common problem in the construction industry and introduced Avvir Inspect. Avvir Inspect uses AI to automatically compare the design model to the point cloud and detect deviations between the two. Avvir Inspect identifies discrepancies between the design intent [BIM] and reality [Point Cloud] to uncover potential issues. Some discrepancies are more critical than others, so Inspect helps parse through the list and highlight critical issues that will clash with future unbuilt elements.
Use cases for point clouds:
Digital Twins with BIM
A common practice is to use the point cloud to create your asset's BIM, a digital illustration of a building's physical and functional characteristics. A BIM is a shared information resource of the building, helping the construction team make critical decisions from planning to handoff during its lifecycle.
The point cloud reflects a new BIM model or updated existing model. Leveraging it and using Avvir helps teams compare As-built conditions of the building against the As-planned to check for conflicts, clashes, and errors.
How it's performed: To generate a BIM from a point cloud, you would use a specialized BIM modeling software to create the model leveraging the collected point cloud data collected.
Some software enables this by manually adding drawings and encoding material information. Other software partially automates the process by identifying objects, surfaces, and even MEP data and using AI to generate a BIM model. These semi-automated processes can be accurate, but they aren't perfect and return false positives and miss important details. A best practice is to always check the work and perform some manual corrections.
The symbiotic relationship means that digital twins can be used like BIM models to review asset information or perform monitoring, simulation, analytics, and control tasks.
How it's performed: Creating a digital twin is a sophisticated, multi-faceted process, so many companies offer services to generate them. The first task is to capture the asset, digitize it, and develop a digital twin that meets specifications and includes the data required to "live" in the model (material data, geometric data, and so forth).
What is the future looking like for point clouds?
At Avvir, we often speculate on the future of construction technologies such as scanning, which is becoming faster, cheaper, and more manageable. For example, we might all have professional-grade scanners on our mobile phones in 5 years. In addition, Apple recently released the Lidar-capable iPhone 12 and new RoomPlan with iPhone 16, enabling users with no previous scanning experience to create point clouds of rooms and objects.
There are many more ways the 3D scanning market could grow and change this year and beyond to elevate the importance of point clouds.
- Expansion of renovation projects and adaptive reuse will accelerate scanning - Using a scanner to create up-to-date building documentation makes the design process more manageable. In addition, it prevents costly errors that result from incorrect information.
- Estimating leverages scanning as a staple in the process - Estimators having access to digital twins can empower their communication assessments with highly accurate data. In addition, leveraging the model brings much more clarity for stakeholders to analyze and discuss costs to reach a winning scenario with high confidence.
- Using scanning for safety compliance increases market share - As scanners become more autonomous and mobile, similar to Boston Dynamic's Spot, it's possible to include the same pattern recognition capabilities. As safety audits become requirements for projects, scanning for safety becomes a capital-investment market segment.
- The expansion of remote work and capabilities continues, scanning for new construction. Although the Covid-19 cloud still lingers, 3D scanning has proven its value throughout construction projects by catching human errors. Using a 3D scanner to verify the site, stakeholders can spot and correct these honest mistakes.
- Mobile phones are accelerating scanning adoption and digital transformation of construction. The future has potential for using mobile phone scanners to create models of the jobsite or detailed-level walk-throughs and clash detections. Having this level of information right at your fingertips democratizes data across the construction lifecycle and increases collaboration on a whole new level.
Innovation awaits the built world.
We are confident that the challenges around point cloud data are temporary. The pace of human innovation – particularly in computer sciences – is accelerating rapidly, so it's simply a matter of time before point clouds can be leveraged to their fullest extent in construction. Avvir will be at the center of innovation using point clouds as the industry continues to evolve.