空间控制技术与应用2023，Vol.49Issue(6)：68-76,9.DOI:10.3969/j.issn.1674-1579.2023.06.007
基于全局注意力机制的单像素成像图像增强方法
Single-Pixel Imaging Enhancement Method Based on Global Attention Mechanism
摘要
Abstract
Single-pixel imaging is an imaging technique that reconstructs a complete image using only non-resolving bucket detectors combined with spatial light modulation information.It features non-local imaging and high sensitiv-ity,making it suitable for ultra-long-distance imaging and detection of non-cooperative targets in outer space.How-ever,it requires multiple spatial light modulations for detection,resulting in low signal-to-noise ratio in the recon-structed images.A global attention mechanism-based image enhancement method for low-sampling rates is presen-ted in this paper.A novel SUNet(swin transformer unet)network is built via Transformer architecture to address the issues of translational invariance and limited global receptive field in traditional convolutional neural networks.Im-proved differential ghost imaging algorithm based on CC(cake-cutting)sequence is employed to reconstruct low-quality images under low sampling conditions,which are then enhanced by SUNet.Experimental results show that,compared to the GIDC(ghost imaging using deep neural network constraint)method proposed in 2022,this ap-proach achieves 3.29 dB improvement in peak signal-to-noise ratio and 8％increase in structural similarity at 0.1 sampling rate,providing a new technological avenue for spatial detection in single-pixel imaging.关键词
单像素成像/全局注意力机制/图像增强/空间探测Key words
single-pixel imaging/global attention mechanism/image enhancement/space exploration分类
信息技术与安全科学引用本文复制引用
刘辉,杨照华,吴云,赵梓栋,余远金..基于全局注意力机制的单像素成像图像增强方法[J].空间控制技术与应用,2023,49(6):68-76,9.基金项目
国家自然科学基金资助项目(61973018)和北京控制工程研究所空间光电测量与感知实验室开放基金资助(LabSOMP-2022-04)Supported by National Science Foundation of China(61973018)and Open Fund of Space Photoelectric Measurement and Sensing Laboratory of Beijing Institute of Control Engineering(LabSOMP-2022-04) （61973018）
