Professor Cho Sung-ho’s research team at the Korea Advanced Institute of Science and Technology (KAIST) has developed an algorithm that creates 3D models of large areas in a short amount of time.
Used on a drones, the technology can conduct land surveys for construction and safety assessments. The researchers created 3D image of the Statue of Liberty to demonstrate the potential.
Introducing autonomy to cars and unlocking well hidden structural secrets
Large area 3D scanning is a valuable technology used in the operation of autonomous vehicles. In some vehicles, self driving capabilities are enabled by Light Detection and Ranging (LiDAR), that effectively serves as the “eyes” of the car.
Large area 3D scanning technologies are also used in structural and conservation projects. The technique can reveal secrets hidden with the structure of a building, as for example in the University of Nottingham’s surveillance of an English Tudor country house.
Though the technology has proved successful in such cases 3D scanning large areas, such as the underground tunnels of Italy’s main cities, still remains a challenge.
Using beams of light, 3D scanners create a map of an area or object by recording the feedback of these beams between the scanned surface and the device. The geometrically complex parts of an area then, with many surfaces, require more effort from the 3D scanner than a comparatively flat or linear surface. This discrepancy is the primary focus of the algorithm developed at KAIST.
Scanning the Statue of Liberty in minutes
The KAIST algorithm doesn’t require any extra information about the to-be-scanned structure prior to flight. Over a large area, the drone scans the surface of a structure at close range to achieve the optimal detail. Instead of simply sweeping over the surface, affording the same time to each point, the KAIST algorithm teaches the drone which areas need extra time to scan.
The device then spends more time on geometrically complex areas during a single workflow. This reduces the number of gaps that occur in the 3D map of a structure, eliminating the need for the drone to return to the same spot at a later time.
In a simulated 3D scan of the Statue of Liberty, the KAIST algorithm took just 30 minutes to create a sample map of the structure. Without the algorithm, the drone would take 51 minutes and 20 seconds to explore the structure. In addition, the technique employs “Next Best View” detection, which allows the drone to predict which part of a wide are it should go to next.
Speaking to Korea’s etnews, KAIST’s Professor Cho Sung-ho comments,
By utilizing drones that freely move anywhere more effectively, we have developed a technology that can computerize information on variety of buildings and structures. This technology will be able to reduce time and cost of establishing information as it can be used for construction industries and industries regarding safety and variety of industries that utilize contents.
For all the latest 3D scanning and 3D printing news subscribe to our free newsletter. If you want to join the conversation then follow the most popular social media accounts in the 3D printing industry.
Featured image: The Korea Advanced Institute of Science and Technology (KASIT) in Daejeon, South Korea. Photo via KAIST