Non-contact hyperspectral technique captures “five-dimensional” images, and with miniaturisation could be housed in smartphones
Engineers in Germany developed the system as a tool for optical sorting of products and precision agriculture. Hyperspectral imaging involves capturing more wavelengths of light that are visible to the human eye; the researchers, from the Friedrich Schiller University Jena and the Fraunhofer Institute for Applied Optics and Precision Engineering, also in Jena, call the technique 5D imaging because as well as capturing the spatial coordinates of the objects to which it is applied, it also captures the precise wavelengths of light reflected from the object from the visible to the near-infrared electromagnetic spectrum and how both variables change in time.
In a paper in the journal Optics Express, team leader Stefan Heist, of the Schiller University, who collaborated with Gunther Notni’s research group from Ilmenau University of Technology, explains how the project culminated in the building of a prototype imager, around the size of a laptop computer. This contains a pair of hyperspectral “snapshot cameras” which are arranged to capture 3D images in the same way that human eyes do. By identifying specific features of the objects that are present in both camera views, the device creates data points capturing the entire surface of the object in x, y and z coordinates. This, however, only works if the object has significant texture or structure. If this is not the case, the device uses a high-speed mechanical projector to illuminate the surface with moving aperiodic light patterns. This allows accurate mapping of the surface features, and hyperspectral reflectance information is then mapped onto the surface.
“Our earlier development of a system projecting aperiodic patterns by a rotating wheel made it possible to project pattern sequences at potentially very high frame rates and outside the visible spectral range,” said Heist. “New hyperspectral snapshot cameras were also an important component because they allow spatially and spectrally resolved information to be captured in a single image, without any scanning.”
“State-of-the-art systems that aim to determine the shape of the objects and their spectral properties are based on multiple sensors, offer low accuracy or require long measurement times,” Heist added. “In contrast, our approach combines excellent spatial and spectral resolution, great depth accuracy and high frame rates in a single compact system.”
To test the system, Heist’s team used it to digitally document a historical relief globe made in 1885. They also mapped the surface of a subject’s hand to demonstrate that the system could identify veins. To demonstrate its usefulness for agricultural applications, they observed a leaf of a citrus plant and showed that the 5D images detected when it was absorbing water.
Further refinements to the equipment could include miniaturising it to fit into a mobile phone. This might allow it to be used for personal medical monitoring, or to detect and measure the ripeness of fruit. Using cameras with a higher signal-to-noise ratio and high imaging rates would allow the system to analyse dynamically changing image properties, and concentrating on certain wavelengths might allow it to detect chemical leaks. Heist also believes the system could be used for security applications, such as identifying people in crowds. What’s more, “because our imaging system doesn’t require contact with the object, it can be used to record historically valuable artifacts or artwork,” he said.