A camera technology was developed that can snap a picture of extremely fast events on the order of a picosecond (11 zeros after the decimal point). Compressed ultrafast photography (CUP) captures the entire process in real time and with unparalleled resolution with just one click. The spatial and temporal information is first compressed into an image and then, using a reconstruction algorithm, it is converted into a video.
Until now, this technique was limited to visible and near-infrared wavelengths and thus to a specific category of physical events. Many phenomena that occur on very short time scales also take place on a very small spatial scale. To see them, one needs to sense shorter wavelengths. Doing this in the ultraviolet (UV) or even X-ray ranges is a step toward that goal.
To record in this new range of wavelengths and develop the technique into a user-friendly product, researchers designed a compact UV-CUP system that features a patterned photocathode, which is used to simultaneously detect and encode “black light. Like a standard camera, the technology is passive — it does not produce light but rather receives it. Therefore, the photocathode had to be sensitive to the photons emitted as UV light. This design makes the technique a stand-alone system that can be easily integrated into various experimental platforms.
Taking the picture, however, is only the first half of the job — it also has to be reconstructed. To do this, the researchers developed a new algorithm that is more efficient than standard algorithms. Rather than solve the reconstruction problem as a lump, the algorithm divides the reconstruction into smaller problems that it tackles individually.
With innovations in both hardware and software, UV-CUP has an imaging speed of 0.5 trillion frames per second. It produces videos with 1,500 frames in large format. As a light-speed imager, UV-CUP sees flying UV photons in real time.
Even though it is the fastest camera in the world, the device can still be improved; for example, the chosen photocathodes have limited efficiency, so researchers want to explore other types of materials. Regarding reconstruction, the algorithm has greatly improved the quality but it still takes quite a long time. Artificial intelligence could be used to improve it.