Using a single-pixel camera operating at THz frequencies, researchers at the University of Sussex have discovered a way to capture with a high degree of accuracy not just the shape of an object, but also its chemical composition. The technique uses time-resolved measurement to reconstruct the complexity of an object. Time-resolved nonlinear ghost imaging. Courtesy of University of Sussex/ACS Photonics. The Sussex team’s approach was to illuminate an object with patterns of THz light containing a broad spectrum of colors. The team then used a single-pixel camera to capture the light that the object reflected for each pattern. The camera was able to detect how the pulse of light was altered in time by the object, even if the THz pulse was an extremely short event. Based on this information and what was known about the shape of the patterns, the researchers were able to derive the shape and composition of the object. The novel methodology, which the researchers describe as nonlinear ghost imaging, theoretically demonstrates potential advantages over state-of-the-art imaging systems for the THz frequency range. “Our approach produces a new type of image which is quite different from what you would get from a standard single-pixel camera as it provides much more information on the object,” said researcher Juan Sebastian Totero Gongora. “Compared to prior single-pixel images, we also demonstrated that our resolution is inherently higher.” “Previous approaches to THz single-pixel cameras cannot preserve the complete information on an object, but we understood where the issue lay and identified a way to extract a more complete image,” said professor Marco Peccianti. “We hope that a similar system to ours could be used in real-life applications in biology, medicine, and security to determine the chemical composition of an object and its spatial distribution in just one step.” In addition to its potential impact on THz cameras, the nonlinear ghost imaging technique could be used to design high-resolution cameras in other frequency ranges, which could be incorporated into technology for collision sensors, body scanners, or ultrafast radars for self-driving cars. The technology was published in ACS Photonics (doi:10.1021/acsphotonics.8b00653).