Keyboard ALT + g to toggle grid overlay
When Jan-Friedrich Köhle wants to survey a property or the route that a new road will take, he simply dons a pair of virtual reality goggles, and in a flash the senior consultant of 5D Institut GmbH from Friedberg in Germany sees a bridge, a slip road, or rolling meadows soon to be crossed by a new bypass before him. With a realistic and immersive experience, the landscapes and structures differ little from their real-life counterparts. “Details just one or two millimeters across can be seen in their true colors and in photographic quality,” Köhle says enthusiastically. The infrastructure expert and his 20-plus-strong team at the 5D Institute – part of the THM, University of Applied Sciences, deployed a drone to survey the route or structure and that they have processed the images photogrammetrically to create a model in 2016.
“The use of drones is particularly promising for inspecting and maintaining bridges,” says Professor Joaquín Díaz, partner and scientific advisor at the 5D Institute. There are roughly 39,000 bridges on German trunk roads, and the state of each one has to be checked about once every two years. Thanks to the technological advancement of drones, organizations like government agencies can now use drones instead of building lifts and scaffolding, saving time and money. Drones and photogrammetric methods allow bridges to be inspected for renovation work up to four times faster with than with laser scanners. “We discovered this when we surveyed a bridge together with a team of surveyors who were using scanners,” Jan-Friedrich Köhle reports. He and his colleagues finished in a single morning, while their scanner-equipped colleagues took two full days.
There are practical uses for drones in road construction as well. Contractors with drones can save up to 30 percent of the costs of building supervision and documentation. “The savings that can potentially be achieved are far greater,” says Professor Díaz. “After all, drones provide documentation of significantly better quality than previous methods.” Until now, the professor of IT in construction says that it has been customary to inspect construction work on site and compare the findings with the plans. In the process, discrepancies are often overlooked. “By contrast, if I fly over the site of a new road once a week and plan the route using a digital model, I get an extremely accurate comparison of the constructed road with the plans,” Professor Díaz explains. “And because the drone doesn’t miss anything, it also provides robust data that can be used to determine quantities and costs.”
“Using Revit and the Dynamo scripts, we can easily transfer data to other systems.”
Before their drones take off, Díaz, Köhle and his colleagues calculate the flight path using a georeferenced Autodesk Revit model. They specify parameters such as the desired resolution of the point cloud and the camera configuration. With the help of a tool developed in Autodesk Dynamo, they determine the optimum recording positions and generate the flight path, taking energy consumption into account. They then load the data into the drone’s control unit.
Communication between the drone and the digital model works by utilizing Autodesk Revit and the Dynamo programming interface. “Using Revit and the Dynamo scripts, we can easily transfer data to other systems,” Köhle explains.
Köhle can then be confident that his drones will survey and measure a building’s facade or a construction site automatically. The aerial survey is an iterative process. After a first rough survey, the software calculates points at which the drone needs to take another photograph in order to achieve the required image quality for photogrammetric processing. “That also enables us to fly around obstacles like trees, whose locations are not noted in the model,” he adds. The photographs, which usually number in the hundreds, as well as the associated geodata are processed by a photogrammetry tool such as Autodesk Forge Reality Capture API or Autodesk ReCap. The result is a point cloud or surface model of extremely high optical quality.
After manual post-processing, the team can transform the surface model into a BIM model. This can then be used to determine the required material quantities and costs. “Based on their dimensions, our design software calculates the volume of the individual components and, with reference to stored cost data and calculation rules, is able to determine the required quantities and costs,” Köhle explains. In this way, specifications for price quotations can be created semi-automatically. Since the 5D Institute links components with working hours, the resulting model not only considers the costs but construction time as well.
“Inaccurate manual calculations, where surveyors obtained dimensions from a paper plan with a ruler and then used those measurements to calculate the required quantities, are now a thing of the past,” Köhle says. Cost drivers such as quantity calculation errors and missing items that must be added later are avoided from the outset.
However, before the automated approach can be used to its full potential, Köhle and his team have to create the required component libraries and calculation rules. In many cases they can use library elements that are in the public domain. “Because Autodesk solutions are widely used in the construction industry, a large number of libraries exist that can be integrated into Revit,” Professor Díaz says. “That’s a huge advantage for us.”
“Because Autodesk solutions are widely used in the construction industry, a large number of libraries exist that can be integrated into Revit. That’s a huge advantage for us.”