Image processing—the procedure of certain operations to extract practical information from images—plays a significant role in the rapid development of modern technology. As an indispensable part, edge detection, which involves computing an image gradient to quantify the magnitude and direction of edges in an image, has considerable applications in medical imaging, face recognition and autonomous vehicles.
Nowadays, facing the rapidly growing demands for fast computation, low energy consumption, and parallel processing, all-optical computation is emerging as a promising avenue for edge detection. To date, several optical differentiators have been proposed for real-time and efficient edge detection, including photonic crystals, photonic chips and metasurfaces. However, those devices and structures are usually fixed once fabricated, and their functionality is always static. The highly limited non-reconfigurable photonic devices cast a shadow on the dynamic tunability of edge-enhanced imaging, thus dampening the potential of this technology.
Overcoming this challenge was the motivation behind the newest National Science Review publication led by the research team of Associate Professor Peng Chen and Professor Yan-Qing Lu from Nanjing University, together with the research team of Associate Professor Wan-Long Zhang and Professor Xiao-Cong Yuan from Shenzhen University. The researchers proposed a fast-switchable optical-edge-detection scheme based on a ferroelectric liquid crystal (FLC) topological structure.
Liquid crystals (LCs) are well known for their good sensitivity to external stimuli such as heat, electric/magnetic fields and light irradiation, and have undergone numerous developments in dynamic functional devices. "As an arresting LC material, ferroelectric chiral smectic C LC is characterized by the in-plane switching of LC directors under an external electric field and the corresponding ultra-fast electro-optical response." Associate Professor Peng Chen, focusing on nanostructured liquid crystals and corresponding optical applications, pointed out. "FLC can be rationally expected to be a strong candidate for the new scheme of dynamic edge-enhanced imaging."
The electrically suppressed helix (ESH) mode stands out due to its tunable optical axis with an ultra-high speed, where the chiral structure of FLCs is highly suppressed by a proper electric field. The team thoroughly investigated the reconfigurable optical axis and resolved the limitation of common binary FLC structures, and propose a fast selective and polychromatic optical edge detection strategy based on inhomogeneously nanostructured FLCs.
The team utilized the photo-alignment technology to build a chiral lamellar structure with azimuthal variation, constructing topological heliconical superstructures in FLCs. Due to the unique mechanism of the ESH mode, the overall space-variant optical axis rotates synchronously with opposite electric field polarities.
"Thanks to the reconfigurable FLC superstructure, nearly orthogonal vector beams can be dynamically generated under the same incidence." Wen Chen, the Ph.D. candidate supervised by Peng Chen and first author of the study, says. "Thus, real-time and high-quality edge-enhanced imaging can be realized based on a conventional 4f system cooperated with crossed polarizers. Horizontal or vertical edges can be selectively highlighted by alternating the polarity of the applied electric field."
"Excitingly, the measured switching time between the horizontal and vertical edge is as fast as 57 μs." says the Ph.D. candidate Dong Zhu, the co-first author of the publication. "The proposed high-speed FLC device provides a fast-switchable solution between orthogonal directional derivative operations, which is promising to work as the controllable processor in optical computing. Moreover, the rapid edge selection might be compatible with commercial high-speed imagers, benefitting the fast acquisition and real-time exhibition of multi-dimensional light information."
In this study, the dynamic functionality remains great over 2000 cycles. Since unmodulated light is filtered by the crossed polarizers here, high-contrast edge imaging is also efficiently achieved throughout an ultra-broad spectrum.
"The excellent reliability and reversibility of the dynamic edge imaging are vividly verified." says the corresponding author and the team leader Prof. Yan-Qing Lu. "This work explores the potential of topological heliconical superstructures in the emerging frontiers of analog optical computing, and discloses their unprecedented capabilities in the fields of machine vision. More possibilities are awaiting us."
