Experimental research on image transmission over one hundred metres with limited bandwidth in a semantic ultraviolet communication system

Currently, ultraviolet (UV) communication is proposed for use in electromagnetic interference environments, while the efficient transmission of images with limited bandwidth has posed a persistent challenge. This paper introduces an light emitting diode-based UV communication system combined with semantic compression for image transmission. The experimental results show that the proposed semantic ultraviolet communication (SUC) system could transmit a 7.19 MB image to a file of approximately 304 KB bandwidth with the aid of deep learning. In addition, we calculated the SSIM (structure similarity index measure) and the PSNR (peak signal-to-noise ratio) to assess the robustness of the SUC system. The experimental results demonstrated that the system could work well over 100 metres while sending images with good quality in the face of error bits interference.