计算光谱成像系统及光谱重建算法

打开文本图片集

关键词：计算成像；光谱成像；压缩感知；深度学习

中图分类号：O433；TP391 文献标识码：A

doi：10.37188/OPE.20263401.0001 CSTR：32169.14.OPE.20263401.0001

Computational spectral imaging systems and reconstruction algorithms

LIU Xinyu 1，2 ，CHEN Yating1，WU Jiachen1，MA Yuchen1，LI Yume 1 ， ZHANG Shuhe 1 ZHENG Zhenrong²，CAO Liangcail*

（1.Department of Precision Instruments，Tsinghua University，Beijing 1OoO84，China; 2. College of Optical Science and Engineering， Zhejiang University， Hangzhou 31O027 ，China） * Corresponding author，E-mail： clc@tsinghua. edu. cn

Abstract： Computational spectral imaging，grounded in compressed sensing theory，incorporates optical encoding elements to project high-dimensional spectral image data into low-dimensional measurements， which are subsequently decoded into spectral images using advanced reconstruction algorithms. This paradigm offers notable advantages in system compactness，acquisition speed，and manufacturing cost. In recent years，rapid progress has been achieved in both theoretical development and system implementation， resulting in a growing body of high-quality research. Concurrently，consumer-oriented deployments have expanded to platforms such as smartphones，unmanned aerial vehicles，and remote-sensing satelites，enabling diverse applications in color imaging，environmental monitoring，and medical diagnostics. In this paper，the theoretical foundations and methodological advances of computational spectral imaging are systematically reviewed. Representative optical encoding strategies-including amplitude encoding， wavelength encoding，wavefront encoding，and multi-aperture encoding-are examined，along with mainstream reconstruction approaches ranging from iterative algorithms with prior constraints to end-to-end deep learning models.Finally，emerging trends and key challenges are discussed. Given its strong relevance to strategic emerging industries，including intelligent manufacturing，artificial intellgence，the low-altitude economy，and smart agriculture，computational spectral imaging is expected to play an increasingly important role across a broad range of applications.

Key words： computational imaging； spectral imaging; compressed sensing; deep learning

1引言

光谱被誉为物质的“指纹”，能够揭示物质对不同波长光的独特响应特性。（剩余50390字）