Bioinspired Reconfigurable Vision Sensor Based on InGaN Semiconductor/Hydrogel Heterostructure for Dynamic-Static Fusion Perception.

Abstract

Intelligent optoelectronic devices capable of simultaneously capturing dynamic motion and static images are crucial for enhancing the multi-functional performance and reliability of machine vision systems. Here, a bio-inspired reconfigurable vision sensor architecture composed of InGaN nanowire/hydrogel heterojunction is proposed, where the electron transport dynamics can be electrically modulated at the heterointerface under closed or open-circuit conditions. This simple operation switch enables rapid transition between photodetection and photosynaptic operation modes. Moreover, inspired by the biological receptor-regulation effect, CoO_x nanoparticle decoration on the nanowires are introduced to improve the heterointerfacial charge transport and chemical reaction processes, effectively enhancing dual-mode optoelectronic performance. Specifically, under photodetection mode, the vision sensor shows an impressive responsivity and microsecond-level response speed, demonstrating fast static image sensing. Under photosynaptic mode, it exhibits typical synaptic behavior, achieving temporal-dependent image memory and enhancement, thereby supporting dynamic motion perception through fusing temporal information. Finally, a chameleon vision-inspired binocular artificial vision system is constructed, which simultaneously recognizes motion trajectories and identifies the absolute positions of moving objects under different motion paths, demonstrating dynamic-static fusion perception capabilities. This work presents a promising architecture for high-performance, self-powered, dual-function vision sensors, paving the way toward next-generation advanced and energy-efficient artificial vision systems.