This makes cameras holding trillions of images per second possible

In some cases, surveillance cameras are used for scientific investigation of processes in physics or chemistry. However, these do not always keep up with the speed of operations – so researchers are forced to resort to tricks and conversions in order to achieve high frame rates. Scientists at the National Institute for Research and Security (INRS) in Canada have set a new record: up to 156.3 terafram per second (TFPS) can be achieved using a hardware arrangement called “SCARF.”

SCARF stands for “Swept Coded Aperture Real-Time Femtophotography” and uses ideas that are already known. The object being examined is illuminated with different colored lights, each representing a single image. Spatial encryption is then performed using an encoding mask that hides specific pixels. For this purpose, the incident light is split into its corresponding spectral components using an optical barrier, creating a time-varying pattern. The split light beam is then combined using another grid, so that the individual images are precisely stitched together again. A specific one ADMM algorithm These individual images can eventually be separated from this pattern and illumination frequency sequence and digitized using a classic CCD sensor.

The research team led by Professor Jinyang Liang Fei INRS website The structure with no additional mechanical components also ensures that SCARF is significantly more economical for such projects, he explains. Quebec-based companies Axis Photonique and Few-Cycle are currently working with scientists to develop a “marketable version” of the patent-pending design.

source: Nature.com via Golem