高级检索

    计算赋能的光学成像:面向轻量化的光学-算法联合设计技术研究

    Computation-Enabled Optical Imaging: A Technical Study of Optico-Algorithmic Co-Design Techniques for Lightweight Optical Systems

    • 摘要: 针对光电设备中光学系统轻量化与高性能难以兼顾的问题,提出一种光学-算法联合设计的简单镜头成像方法。通过二次函数自适应波段权重优化校正中心波段像差,并利用交叉通道先验将高频信息传递至其余波段,实现图像复原;同时提出一种适用于简单镜头的全光谱点扩散函数高效测量方法,支撑复原算法。基于所提方法设计一款F数为6、焦距50 mm的双分离简单镜头,实验表明其成像与同参数标准镜头的结构相似度指数达0.96,在保证成像质量的同时,透镜数量减少2片,总重量降低约59%。结果表明,该方法可在基本保持成像性能的前提下显著实现系统轻量化,对光学系统的轻量化、便携化发展具有重要推动意义。

       

      Abstract: To adress the longstanding challenge of simultaneously achieving lightweight design and high imaging performance in the optical systems of optoelectronic equipment, a novel lens imaging framework grounded in integrated optical-algorithmic co-design is proposed in this paper. Central-band aberrations are corrected via quadratic-function-based adaptive band-weight optimization, while high-frequency detail preservation across spectral bands is enabled through cross-channel prior exploitation. Furthermore, an efficient full-spectrum point spread function (PSF) measurement technique tailored for simple-lens configurations is introduced, which directly supports the proposed restoration pipeline. Leveraging this framework, a dual-separation simple lens with an F-number of 6 and a focal length of 50 mm is designed. Experimental evaluation demonstrates that the prototype achieves imaging fidelity comparable to that of a conventional multi-element lens of identical specifications—quantified by a structural similarity (SSIM) index of 0.96. Critically, the design reduces the total number of lens elements by two and achieves a 59% reduction in overall mass. These results substantiate that the proposed approach enables substantial system miniaturization and weight reduction without compromising optical performance, offering tangible advancement toward lightweight, portable, and cost-effective optoelectronic systems.

       

    /

    返回文章
    返回