In recent years, construction companies have started to use drones as mapping and surveying tools, replacing traditional week-long ground surveys with just a few hours in the air. Aerial mapping of construction sites makes it easier to continually monitor construction, which in turn makes it easier to spot mistakes before they become expensive or difficult-to-fix.
In a recent report, Mapping the Friendly Skies: How Drone Mapping is Changing the Construction Industry, engineering.com interviewed professionals from across the construction industry to talk about drone mapping best-practices, and the factors that construction professionals should consider before buying their own drones.
“In the construction space, drones are very disruptive, because, for the first time ever, they have made it cost-effective to get huge amounts of data on what’s happening on the construction site,” said Tristan Randall, strategic projects executive at Autodesk.
In the past five years, drone have made the jump from unreliable toy to hardworking tool. Today's mapping drones don't fly away, crash unexpectedly, or struggle to carry light loads or stay in the air. They also cost significantly less than they used to, putting industrial drone ownership within the reach of smaller construction companies. The report finds that drone use on construction sites has jumped significantly in the past few years, and that it will likely continue to increase significantly in the next few years as well.
The report also probes the different kinds of drones, imaging systems, and mapping software available to construction professionals. In the construction industry, multirotor drones (drones with more than one "propeller") are the most common, because they're inexpensive to fly, easy to position, and least likely to cause damage or injury. Quadcopter (four-propeller) drones are the most common because they’re relatively inexpensive, but hexcopters and octocopters are also useful on construction sites and offer greater stability.
There are two main types of drone imaging systems, photogrammetry and LiDAR, and both systems are useful for different projects. Photogrammetry involves taking pictures of the site from different angles to build a 3D RGB model, whereas LiDAR (Light Detection and Ranging) involves firing multiple rapid pulses of laser light toward the ground and building a point cloud depth map from the time it takes for the beams to return.
Photogrammetry is more common, because it was initially easier to mount a camera on a drone than it was to mount a LiDAR scanner. LiDAR solutions are still in their infancy, and generally need to be custom-built (with the exception of the new FARO/Stormbee collaboration), but they present exciting possibilities, and can create a more precise dataset. Because LiDAR scans have a built-in measurement system, the construction team doesn't need to place as many ground control points to ensure that the scan is accurate, saving a significant amount of time. LiDAR scanners can also make reliable maps under difficult lighting conditions, because they don't rely on ambient light like photogrammetry does.
With the rise of aerial data mapping, more and more mainstream BIM companies like Bentley, Autodesk, and Trimble have started offering ways to use context capture data directly in BIM applications. Applications like Bentley’s ContextCapture let users make usable reality meshes out of raw photogrammetry or LiDAR data, whereas software like FARO’s BuildIT Construction lets users compare the reality capture data to the initial BIM model, letting them easily screen for deviances or mistakes.
For these companies, the goal is to offer an end-to-end solution that makes context-capture data a natural part of the user's workflow. “With our construction BIM business, we are really focusing on what we call the traceable construction life cycle,” said Randall. “We want to offer products along the whole construction life cycle to make the construction traceable, reproducible and documented.”
One important tool in the drone user's arsenal is the increasing number of cloud-based services to store and analyze the data. Drone imaging, especially if it's done via LiDAR, creates fast datasets that can be expensive to store. Companies like 3DR and Propeller have started offering cloud services that also include built-in data analytics and image processing. These types of cloud storage are additionally useful because they mean that project stakeholder can access data on multiple different devices.
“There are a lot of ways to process drone data, and one of the most traditional ways is to process it through very expensive desktop programs,” said Rory San Miguel, cofounder of Propeller Aeropoints, a drone software provider. “What we’ve done is we’ve put that in the cloud, and we’ve paired that alongside photogrammetry experts to do that work for you.”
Another recent development in the field is using AI and machine learning to analyze drone datasets. Machine learning is used most often in inspection and maintenance surveys, where the user needs to look through vast datasets of similar-looking terrain to find defects or damage. To "train" the algorithms, users just need to feed them a large dataset with the types of flaws it needs to look for.
Drone users and drone creators alike believe that the field is moving in the direction of greater automation and easier use. Christian Sanz, CEO of specialized machine learning company Skycatch, said, “Automation in construction is no longer something to look out for—four or five years in the future. When you go to a job site, you should expect to see robots on the ground.”
Read the full report: Mapping the Friendly Skies: How Drone Mapping is Changing the Construction Industry