Spectrum-dependent image convolutional processing via a 2D polarization-sensitive photodetector

Abstract

Abstract With the advancement of photodetectors and the widespread deployment of high-definition image sensors, the growth in processing capacity of traditional processors significantly lags behind the demands of real-time image processing. Consequently, achieving real-time image processing at the sensor level has emerged as a prominent research domain in the photodetector field, holding potential applications in various fields including remote sensing, autonomous driving, and road monitoring. This advancement underscores the need for photodetectors with enhanced multifunctional integration capabilities, capable of performing real-time computations on optical or electrical signals. In this study, we employ an innovative two-dimensional p-type semiconductor GaTe0.5Se0.5 as the light absorption layer to construct a photodetector with a polarization-sensitive wide-spectral photoresponse within the wavelength range of 255 to 1064 nm. This design enables the realization of a photodetector that integrates real-time sensing and computation. Leveraging the wide-spectral photoresponse of the photodetector, wide-spectral imaging was realized within the wavelength range of 390–810 nm during experimentation. Furthermore, by utilizing the tunable photoresponses of the photodetector through varying polarization angles, real-time image convolutional processing is enabled, aided by configuring appropriate convolution kernels based on the polarization-sensitive photoresponse values. Our proof-of-concept device, featuring spectrum-dependent polarization-sensitive photoresponses across a wide-spectral range, successfully accomplishes real-time image convolutional processing, including sharpness and edge extraction operations, thereby facilitating more precise capture of image details. The innovative design of the polarization-sensitive wide-spectral GaTe0.5Se0.5-based photodetector constitutes a notable contribution to the domain of real-time image perception and processing.