【作者】 刘洁；

【导师】 欧海燕；

【作者基本信息】 电子科技大学 ， 电子信息（专业学位）， 2025， 硕士

【摘要】 光学扫描全息技术作为一种特殊的数字全息技术,通过主动扫描将三维物体的振幅和相位信息完整记录于二维数字全息图中。该技术凭借其实时性、高分辨率和非侵入性等优势,已成为国内外研究热点,在生物医学、微纳器件检测及遥感等领域展现出广阔的应用前景。在光学扫描全息技术的应用中,全息图重建是获取物体清晰图像的关键步骤,而轴向定位的精确性直接影响重建图像的清晰度和聚焦效果。不准确的轴向定位可能导致图像模糊或失焦,进而影响后续分析与应用。针对这一问题,本文基于光学扫描全息记录与重建原理,重点研究轴向定位问题,通过理论分析、仿真和实验验证,深入探讨相关自聚焦算法。主要研究内容包括以下几个方面:(1)从波动方程、菲涅尔衍射定理等基本光学理论出发,结合信号处理方法,推导了光学扫描全息的数学模型,并分析了光学扫描全息重建过程中所面临的定位问题。(2)在传统时间反演与多信号分类(Time-reversal and Multiple Signal Classi-fication,TR-MUSIC)自聚焦算法的基础上,提出了一种基于李萨如扫描的改进TR-MUSIC算法,以优化分辨率与计算效率的平衡。通过MATLAB仿真和实验验证,改进的算法在定位精度、计算效率、分辨率和抗噪声性能方面均表现出显著优势,峰值突出度从0.21增加到0.34。(3)提出了一种基于谷本系数(Tanimoto Coefficient,TC)的高效自聚焦算法。该算法通过分离全息图的实部与虚部并分别重建,利用谷本系数评估重建图像的相似度,实现精确轴向定位。结合连通域方法,有效抑制了离焦噪声的干扰。不同于现有基于相似性的自聚焦方法,该双通道TC方法无需依赖RGB(Red,Green,Blue)信息。同时,相比TR-MUSIC方法,计算速度提升超过90倍。MATLAB仿真与实验结果均验证了该算法的高效性与可靠性。更多还原

【Abstract】 Optical scanning holography(OSH),as an special digital holographic technique,em-ploys active scanning to comprehensively record the amplitude and phase information of three-dimensional(3D)objects into two-dimensional(2D)digital holograms.Owing to its advantages such as real-time capability,high resolution,and non-invasiveness,this technology has become a focal point of research both domestically and internationally,demonstrating vast application potential in fields such as biomedicine,micro-nano de-vice inspection,and remote sensing.In the application of OSH,hologram reconstruction is a critical step for obtaining clear images of objects,and the accuracy of axial positioning directly impacts the clarity and focus of the reconstructed images.Inaccurate axial positioning may result in blurred or defocused images,thereby affecting subsequent analysis and application.To address this issue,this thesis focuses on the axial positioning problem based on the principles of optical scanning holographic recording and reconstruction,and conducts an in-depth exploration of related autofocusing algorithms through theoretical analysis,simulation,and experimental validation.The main research contents include the following aspects:(1)Based on fundamental optical theories such as the wave equation and Fresnel diffraction theorem,combined with signal processing methods,the mathematical model of dual-pupil OSH was derived,and the localization challenges in the optical scanning holographic reconstruction process were analyzed.(2)Building upon the traditional time-reversal and multiple signal classification(TR-MUSIC)autofocusing algorithm,an improved TR-MUSIC algorithm incorporating Lis-sajous scanning is proposed to optimize the trade-off between resolution and computa-tional efficiency.MATLAB simulations and experimental validations demonstrate that the improved algorithm exhibits significant advantages in localization accuracy,compu-tational efficiency,resolution,and noise robustness,with the peak prominence increasing from 0.21 to 0.34.(3)An efficient autofocusing algorithm based on the Tanimoto coefficient(TC)is proposed.This algorithm reconstructs the real and imaginary parts of the hologram sepa-rately and evaluates the similarity of the reconstructed images using the Tanimoto coeffi-cient to achieve precise axial localization.By integrating the connected domain method,the algorithm effectively suppresses defocus noise.Unlike existing similarity-based aut-ofocusing methods,this dual-channel TC approach does not rely on(Red,Green,Blue)RGB information.Additionally,compared to the TR-MUSIC method,it improves com-putational speed by more than 90 times.MATLAB simulations and experimental results verify the efficiency and reliability of the proposed algorithm.更多还原