Hyperspectral imaging is a technique combining imaging and spectroscopy resulting in images with extraordinary precision and detail. Current approaches to capture hyperspectral images are costly and time-consuming. The proposed technique makes use of inexpensive filters and reduces the number of required exposures, thereby improving the efficiency and practicability of obtaining hyperspectral images.
Unlike traditional photographs, hyperspectral images capture a large number of contiguous spectral bands at each pixel, which can be used to characterize the objects in an image. Currently, hyperspectral imaging requires capturing many (~10-20) individual exposures with a different, expensive filter used for each exposure. The need for specialized and costly filters coupled with the time-consuming approach limit the applications for obtaining these high-quality images. UCI researchers developed a technology for hyperspectral imaging where only 3 exposures are needed using 2 inexpensive filters.
In the proposed technology, the spectral information is encoded in a highly efficient manner using a “phasor” technique, allowing for hyperspectral image files smaller than traditional encoding methods. The spectral phasor is calculated with a simple calibration from only 3 exposures using a standard monochrome camera. This time- and cost-saving technique improves the practicality of hyperspectral imaging for many applications.
A prototype for this technology has been demonstrated
Alexander Dvornikov and Enrico Gratton, "Hyperspectral imaging in highly scattering media by the spectral phasor approach using two filters," Biomed. Opt. Express 9, 3503-3511 (2018) Link