Ever wonder what the raised height of Washington’s head is on a U.S. quarter? Or the thickness of different inks on a chocolate bar wrapper?
Researchers at Carnegie Mellon University’s School of Computer Science and the Israel Institute of Technology’s Technion have created a new generation of 3D sensors capable of capturing such micron-scale depth detail.
“The sensors we have been working on have a resolution measured at the micrometer scale,” said Ioannis Gkioulekas, an assistant professor in CMU’s Robotics Institute. “That is a thousand times more resolution than current LiDAR and time-of-flight sensors.”
This incredible amount of depth detail is obtained using a technique called “interferometry,” which measures the interference of superimposed waves — in this case, light — to extract information.
Gkioulekas and Alankar Kotwal, a recent Ph.D. graduate in robotics, will present their work at the Computer Vision and Pattern Recognition Conference this month in Vancouver, Canada.
“One of the examples we spotlight is inspecting the fuselage of a Boeing 747 for small marks from particle impact during flight,” Gkioulekas said. “Other applications we discuss are in fabrication, medical imaging and robotic manipulation.”
The researchers have shown that this type of 3D sensing can be done passively, using only the sun as a light source with no need for lasers or other active lighting. With such minimal lighting requirements, the sensors can run in uncontrolled conditions outdoors using only a small cellphone battery to power the camera.
Detailed imaging is critical for applications such as inspection and fabrication, where fast and accurate 3D sensing at these scales is crucial. It is also valuable in areas like virtual reality, autonomous systems, medical imaging and other computer vision applications that require accurate 3D representations.
Read more about Passive Micron-Scale Time-of-Flight With Sunlight Interferometry and Swept Angle Synthetic Wavelength Interferometry on the Imaging@CMU website.