Overcoming High Bit Error Rate With Wavelength Bipolar Control: A Self‐Powered Ultraviolet Optoelectronic Platform for Underwater Intelligent Imaging and Communication

ABSTRACT Underwater optical communication and intelligent imaging face challenges from ambient light interference and elevated bit error rates in conventional visible‐band systems. To address it, this study presents a self‐powered ultraviolet (UV) optoelectronic platform based on a CuI@ α ‐Ga 2 O 3 nanorod arrays heterojunction to explore its anti‐interference performance in communication. The design integrates the intrinsic solar‐blind spectral selectivity of α ‐Ga 2 O 3 with the high hole mobility of p ‐type CuI to establish a competitive charge‐transfer regime at the semiconductor/electrolyte interfaces. Under zero bias, the device exhibits a unique wavelength‐discriminative bipolar photoresponse: 254 nm illumination excites α ‐Ga 2 O 3 to generate a positive photocurrent, whereas 365 nm light activates CuI to yield a negative photocurrent. This hardware‐level signal polarity inversion directly enables a wavelength shift keying modulation protocol, suppressing the bit error rate and enhancing signal fidelity. Furthermore, by coupling the bipolar photoresponse with a YOLOv11‐based neural network and ASCII encoding, the system achieves high‐precision underwater pattern recognition and intelligent imaging, delivering exceptional recognition accuracy alongside an ultralow bit error rate in UV communication. The wavelength‐dependent polarity inversion increases the separation between logic states and supports high‐fidelity signal transmission. This work enables self‐powered underwater devices, with clear applications in future secure communications and environmental monitoring.