物理与工程2026，Vol.46Issue(5)：70-78,9.DOI:10.27024/j.wlygc.2026.06.24.02
面向AI算力提升物理机制的微透镜光学卷积实验研究
MICROLENS OPTICAL CONVOLUTION FOR PHYSICAL MECHANISMS OF AI COMPUTING POWER ENHANCEMENT
摘要
Abstract
Convolutional neural networks are widely used in tasks such as image recognition due to their excellent performance in two-dimensional data processing.To address the limited AI computing power under the traditional digital integrated circuit computing architecture,this study designs an optical convolution system based on a microlens array.Utilizing the ul-tra-high-speed and high-parallelism characteristics of optical computing,it achieves convolu-tion computation based on the principles of geometrical optics.The system loads image infor-mation and convolution kernels using a spatial light modulator,segments the light field through a microlens array,modulates the segmented sub-regions in parallel using the convolu-tion kernel,and finally converges through the imaging optical path to achieve summation,thus completing the convolution computation in optical space.This study presents the imple-mentation principle of the optical convolution system based on a microlens array,designs the experimental optical path,and proposes implementation methods for key technologies such as system calibration,convolution kernel implementation,and image post-processing.Experi-mental results show that the system possesses basic imaging performance and the ability to implement complex feature operators such as Gaussian operator,Gradient operator,and Laplacian operator.Limited by device accuracy,the system still has errors caused by aberra-tions and diffraction.It can provide verification for the feasibility of optical convolution schemes based on geometrical optical architecture in image preprocessing tasks.关键词
光学信息处理/卷积神经网络/微透镜阵列/图像处理Key words
optical information processing/convolutional neural network/microlens array/image processing引用本文复制引用
王泽生,孟飞扬,杨明睿,林志浩,刘俊衔,孙文博..面向AI算力提升物理机制的微透镜光学卷积实验研究[J].物理与工程,2026,46(5):70-78,9.基金项目
教育部拔尖计划2.0研究课题(20211007),教育部产学合作协同育人项目(231001318115817),清华大学实验室创新基金(202504303). （20211007）
