Costly laser scanners could soon be replaced by the ubiquitous smartphone as the most effective means of performing reality modelling of building and construction sites.
According to Pascal Martinez, director of business development for Bentley Systems, the latest reality capture software can produce high-resolution 3D models of landscapes and buildings using nothing more than multiple photos shot on site.
“This new technology means that you can turn photos into any kind of 3D model or 3D mesh – you can go anywhere on-site and shoot pictures around specific parts of your project in order to transform it into a 3D model,” said Martinez. “If you want more detail and accuracy, it’s just a matter of taking more photos and getting in closer. Zoom with your feet, get closer to the object and shoot more pictures.”
This new form of context capture originally developed by Acute3D possesses a number of advantages compared to conventional laser scanning methods. Chief amongst them is the ability to use devices as cheap and commonplace as smartphones for reality modelling purposes.
“The big advantage of our new method over laser scanners is that it can be performed without costly equipment, simply by using your smartphone or a compact camera,” said Martinez. “These days everybody knows how to use a smartphone as a camera, so everybody virtually has a scanner in his or her pocket.”
The software used to transform a collection of photos taken of a single object or environment into coherent and accurate 3D models also has the added advantage of producing structured meshes as their final product, as opposed to more ambiguous point clouds.
“With laser scanning, what you get is just point clouds, and it’s difficult for the brain to reconcile those points into a 3D object, because we don’t see the world in point clouds – we see the world with surface and texture,” said Martinez. “With meshes you have texture, and texture gives you the real conditions of a physical context or object.
“That’s another level of information that is provided by our software to the human brain, which is important because we don’t produce software for computers – we produce software for human users.”
Another major advantage of photo-based reality modelling is scalability and associated levels of precision. The technology can be used to create a 3D model of a simple object using just a dozen photos, or produce a 3D replica of a city using many thousands or millions of shots.
“Our key differentiator on the market is that we can scale from one object to an entire city using the same technology without any limits,” said Martinez.
“Scalability is what unlocks precision as well. If you want to capture the very precise details of a structure, if you want to inspect the cracks in a bridge or a dam, it’s just a matter of taking more photos and getting in closer in order to incorporate these features into the final 3D model.”
The scalability of photo-based reality capture methods have already been amply demonstrated in North America, where AEROMetrex and Bentley System used the technology to produce a highly detailed, photo-realistic 3D mesh model of the city of Philadelphia in preparation for Pope Francis’ visit in September.
Martinez expects the new technology to be particularly effective for mapping, surveying and inspection purposes when employed in tandem with unmanned aerial vehicles (UAVs), particularly given that the constituent photos used to produce the 3D models lend themselves to scrutiny as well.
“When the time comes for inspection of a building, you can use a drone at an altitude of two hundred metres to take photos of it in order to produce a 3D model with full rendering of all aspects,” he said. “But people can also inspect each of the individual pictures just by looking at or clicking on the 3D model the user can say show me all the pictures that involves this very specific point in full resolution.”