Search
Menu
Spectrogon US - Optical Filters 2024 LB

Motion Blur Correction Method Will Bring High-Quality Imaging to Dark Environments

Facebook X LinkedIn Email
TOKYO, June 9, 2021 — Researchers from Tokyo University of Science introduced a technique that addresses a limitation of single-photon imaging: motion. A team led by professor Takayuki Hamamoto developed an algorithm capable of addressing the blurring caused by the motion of an imaged object, as well as common blurring of the entire image such as that caused by the shaking of the camera.

When taking a picture with a CMOS camera (like the camera found in a smartphone), a moving object can be imaged by reducing the exposure time. Images taken with single-photon cameras (single-photon imaging) are constructed from a series of very short individual exposures that capture consecutive frames. The frames are binary — a grid of ones and zeros that represent whether a photon arrived at each pixel or not during exposure.

The speed at which single-photon imaging can occur, however, is much higher than with a CMOS camera. Due to its entirely digital nature, single-photon imaging allows for clever reconstruction algorithms that can make up for technical limitations or difficult scenarios.
Tokyo University researchers developed a method of deblurring single-photon images containing multiple dynamic moving objects. According to the team, images such as this one, taken with  an ordinary digital camera, may one day see benefit from such processing methods. Courtesy of Pixabay.
Tokyo University researchers developed a method of deblurring single-photon images containing multiple dynamic moving objects. According to the team, images such as this one, taken with an ordinary digital camera, may one day see benefit from such processing methods. Courtesy of Pixabay.


The team’s approach addresses many limitations of existing deblurring techniques for single-photon imaging, which produce low-quality pictures when multiple moving objects are present in the image scene and are moving at different speeds and/or are overlapping each other. Rather than adjusting the entire image according to the estimated motion of a single object, or on the basis of spatial regions where the object is considered to be moving, the method uses a more dynamic strategy. A motion estimation algorithm tracks the movement of individual pixels through statistical evaluations of how bit values change over time (over different bit planes). In this way, experiments showed, the motion of individual objects can be accurately estimated.

Optimax Systems, Inc. - Ultrafast Coatings 2024 MR

“Our tests show that the proposed motion estimation technique produced results with errors of less than one pixel, even in dark conditions with few incident photons,” Hamamoto said.

After obtaining information from the initial algorithm, a second deblurring algorithm groups together pixels with similar motion, identifying separate objects moving at different speeds in each bit plane. Each region of the image can therefore be deblurred independently according to the motions of objects that pass through it. Simulations showed crisp and high-quality images, even in low-light dynamic scenes containing multiple objects moving at various speeds.

The results demonstrate the utility of image processing techniques in improving single-photon image quality.

“Methods for obtaining crisp images in photon-limited situations would be useful in several fields, including medicine, security, and science,” Hamamoto said. “Our approach will hopefully lead to new technology for high-quality imaging in dark environments, like outer space, and super-slow recording that will far exceed the capabilities of today’s fastest cameras.”

Hamamoto noted that even consumer-grade cameras could benefit from progress in single-photon image processing.

The research was published in IEEE Access (www.doi.org/10.1109/ACCESS.2021.3059293).

Published: June 2021
Research & TechnologyImagingsingle-photonsingle-photon imagingsingle photon imagingblurimage blurimage processingimage processing algorithmTokyo UniversityTakayuki Hamamotolow lightmotion blurdeblurringdeblurring algorithmsIEEE Access

We use cookies to improve user experience and analyze our website traffic as stated in our Privacy Policy. By using this website, you agree to the use of cookies unless you have disabled them.