Abstract
Optical technology may provide important architectures for future computing, such as analog optical computing and image processing. Compared with traditional electric operation, optical operation has shown some unique advantages including faster operating speeds and lower power consumption. Here, we propose an optical full differentiator based on the spin–orbit interaction of light at a simple optical interface. The optical differential operation is independent of the wavelength due to the purely geometric nature of the phenomenon. As an important application of the fully differential operation, the wavelength-independent image processing of edge detection is demonstrated. By adjusting the polarization of the incident beam, the one-dimension edge imaging at any desirable direction can be obtained. The wavelength-independent image processing of edge detection provides possible applications in autonomous driving, target recognition, microscopic imaging, and augmented reality.
Highlights
Wavelength-independent optical fully differential operation based on the spin–orbit interaction of light
We have realized an optical fully differential operation based on the spin–orbit interaction of light at an optical interface
This differentiator can calculate the spatial difference of the input wave function in the process of light reflection and has the advantages of superspeed and low power consumption
Summary
In recent years, compared with the maturity of digital circuit technology, there has been a lot of optimism about the future development of optical analog computing. While digital signal processors provide high speed and reliable operation, it has some obvious disadvantages, such as high-power consumption, expensive analog-to-digital converters, and sharp performance degradation at high frequencies. Considering these limitations, it is neither rational nor economical to use digital signal processors to perform specific, simple computing tasks such as differential or integral, equation solving, matrix inversion, edge detection, and image processing. Analog signal processors, due to their wave-based characteristics, potentially have superior performance over digital versions, including faster operating speeds and lower power consumption. Based on the above advantages, the optical differentiator is used to realize ultra-fast parallel image processing, realtime boundary detection, and judgment of boundary direction.. While digital signal processors provide high speed and reliable operation, it has some obvious disadvantages, such as high-power consumption, expensive analog-to-digital converters, and sharp performance degradation at high frequencies.3 Considering these limitations, it is neither rational nor economical to use digital signal processors to perform specific, simple computing tasks such as differential or integral, equation solving, matrix inversion, edge detection, and image processing.. The design of micro–nanostructure has been shown to be extremely complex, so it is very difficult to fabricate these devices accurately and realize optical simulation operation In response to this challenge, optical differentiators based on surface plasmon, photonic crystals, spiral phase contrast, the grating nanostructure, photonic spin Hall effect, geometric spin Hall effect, and metasurface have been developed. The differentiator is based on the principle of spin–orbit interaction of light, and we have demonstrated that its working performance has no wavelength limit
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
